We used a spatially explicit regression model to relate sardine (Sardinops sagax) larval abundance to water column stability, phytoplankton, and zooplankton off South Australia. The distribution of sardine larvae was significantly associated with stability (p , 0.05) and phytoplankton fluorescence or zooplankton displacement volume (p , 0.001) at broad scales. In contrast there was no relationship between sardine larvae, stability, phytoplankton, or zooplankton at medium or fine scales. The relationships are shown to be scale-dependent. Although the results generally support Lasker's (1978) ''stable ocean'' hypothesis, mid-range rather than high stability appears to be favored, in accord with theoretical expectations at both large and small scales.Many studies have attempted to relate recruitment success in clupeoids to environmental and ecological factors. No single factor explains the observed variability in clupeoid recruitment but the role of water column stability is considered to be critical for survival of the early larval stages. Lasker's original idea (Lasker 1978) now referred to as the ''stable ocean'' hypothesis, suggested that survival of first-feeding anchovy larvae and subsequent recruitment would be improved when food of a suitable size and concentration aggregated in layers, associated with a stable water column, in the absence of intense upwelling advecting the larvae into possibly unfavorable food environments. However, evidence for the correlation of stability with larval survival is variable and has seldom been shown to agree with theoretical expectations (Peterman and Bradford 1987;Cury and Roy 1989;Butler 1991). The survival of newly hatched larvae is most likely determined by a combination of physical and biological characteristics, rather than by feeding conditions alone (Muelbert et al. 1994). However, sardine recruitment appears to be determined by the mortality experienced by post-larvae to juvenile (age 0) fish (Watanabe et al. 1996), rather than by the first feeding stage.Stability exerts an effect at a range of scales by altering the prey environment experienced by larval clupeoids as they lose their yolk sacs and begin feeding. In the case of anchovies, the prey are often dinoflagellates (Lasker 1975), whereas first feeding sardine consume copepod nauplii (Watanabe et al. 1998). First feeding larvae must encounter prey of suitable size at sufficient densities to meet their metabolic requirements for growth (Lasker 1975). Aggregation of the prey in layers provides a sufficiently concentrated food source for the larvae to be able to meet their metabolic requirements for growth and to grow rapidly enough to minimize predation mortality. Some species of early stage larval fish can swim well enough to locate the layers of prey and to successfully capture food particles if the stability conditions are favorable (MacKenzie and Kiorboe 2000). Although there is little swimming speed data for sardine larvae, it is thought that postflexion larval stages are likely to be able to move ...