Long-term population studies provide the foundation against which ecological hypotheses on mechanisms of population regulation can be accepted or refuted. One of the classic long-term population studies has focused on the winter moth, Operophtera brumata (L.) (Lepidoptera: Geometridae) and its associated natural enemies. Detailed population studies have been carried out in a native area at Wytham Wood, Oxfordshire (Varley & Gradwell 1963, 1968; Varley, Gradwell & Hassell 1973) and in regions where the moth has been introduced into Canada (Embree 1965, 1991; Roland 1986, 1988). Different regulatory mechanisms have emerged from these studies. At Wytham Wood, mortality occurring in the soil, mainly due to pupal predators, acts in a density-dependent manner and seems to be the primary regulatory process (Varley et al. 1973). In Canada, however, pupal predation seemed much less important and an introduced tachinid parasitoid, Cyzenis albicans (Fall.) (Diptera: Tachinidae), was identified as the main density-dependent process (Embree 1966). More recently, Roland (1986, 1988 1990, 1994) has presented yet further interesting interpretations of the winter moth population dynamics based on his studies at Mount Tolmie Park, British Columbia, Canada. In his most recent paper (Roland 1994), data are presented on the relationships between mortality and the estimated density at different stages in the life cycle between 1982 and 1990 (Fig. 1). During this period the winter moth population declined from initial high levels to relatively low levels (for the final 7 years). In order to determine the importance of these mortalities to the winter moth dynamics, Roland selected only the final six 'equilibrium' years of data for his analysis (Fig. 2). On the basis of this, he concluded that egg and larval mortality (k fec and klarv) were not densitydependent and therefore not important in regulating the winter moth population. Pupal predation by ground beetles, however, did appear to be strongly density-dependent and was thus inferred as the primary cause of stability in the British Columbia winter moth population. In this note, we question this conclusion. In particular, we caution against looking for specific densitydependent processes from relatively stationary populations. By considering all the available data (198290), we show that egg and larval mortality (k fec and klarv), in contrast to pupal mortality (kpred), are densitydependent and could have a major impact on the winter moth population regulation.