AbstractThe efficacy of biological control agents depends on the amount of mortality they impose on pests in the field. Quantifying their impact on a pest is critical for guiding decision‐making in its management. Here, we quantify the relation of population levels of natural unmanipulated populations of a pest to mortality rates imposed by naturally occurring enemies. Mortality beyond first instar caused by natural enemies of Aonidiella aurantii (Maskell) (Hemiptera: Diaspididae) on fruit was studied in four unsprayed citrus orchards in a warm temperate coastal environment in New South Wales, Australia, over approximately three pest generations in the course of a year. The percentage of hosts on fruit suffering parasitism and predation was usually independent of host density. There was no consistent evidence that one parasitoid influenced another, nor that parasitism and predation rates were related. At the orchard scale, it was apparent that parasitism and predation over a time series were related to current or previous host numbers and could become high enough in relation to fecundity rates to cause pest numbers to decline. Temporal feedback to density could be causing the pest populations to fluctuate between limits at economically sustainable levels given additional mortality of the first instars due to expected rates of dispersal and subsequent pre‐ and post‐settlement failures.