The spatial dispersion of armored scale insects; greedy scale,Hemiberlesia rapax (Comstock); and latania scale,Hemiberlesia lataniae (Signoret), was investigated onkiwifruit, Actinidia deliciosa (A. Chevalier) C. F.Liang et A. R. Ferguson, leaves in New Zealand. A universaldescription for dispersion was determined using Taylor’s power law,which encompassed a wide range of different orchards, blocks, blocksizes, sampling times, scale control practices, regions and seasons.Scale density significantly altered dispersion, especially at the highdensities found on unsprayed kiwifruit. Most commercially managedkiwifruit blocks had low densities of <0.5 scale per leaf and had aslightly aggregated scale dispersion. Wilson and Room’s binomialmodel, which incorporates a clumping pattern as a function of density,gave a significant relationship between the proportion of infestedleaves and scale density. The optimal leaf sample sizes were estimatedfor predetermined levels of sampling reliability. Where populationestimates require a high degree of precision and enumerative samplingmethods are used, 2,500 leaves should be sampled when scale densitiesare near the current spray threshold of 4% infested leaves and 500leaves at 20% infested leaves. For management-decision sampling, wherea lower level of precision was acceptable, enumerative sampling wouldrequire that 400 leaves be sampled at 4%; or 85 leaves at 20%infested leaves. With binomial sampling to achieve an equivalent levelof precision an increased sample size of 6–11% is required.