Nematode damage functions: the problems of experimental and sampling error.

The development and use of pest damage functions involves measurement and experimental errors associated with cultural, environmental, and distributional factors. Damage predictions are more valuable if considered with associated probability. Collapsing population densities into a geometric series of population classes allows a pseudo-replication removal of experimental and sampling error in damage function development. Recognition of the nature of sampling error for aggregated populations allows assessment of probability associated with the population estimate. The product of the probabilities incorporated in the damage function and in the population estimate provides a basis for risk analysis of the yield loss prediction and the ensuing management decision.