AbstractSeveral different intertime dissimilarity measures, used with a variety of standardisations of data, were applied to transformed data of macrobenthic species at six sites. These data were obtained over twenty‐two times at 8‐week intervals. Results were evaluated by stepwise regressions (linear, annual cyclical, longer term cycles). Manhattan metric flux between consecutive times was preferred for most subsequent purposes. This flux divided into positive and negative components which measure, summated over all species, rates of recruitment and depletion. Use of algebraic and absolute sums permits evaluation of an index of homeostasis which measures the counterbalancing of recruitment by some species and depletion by others; this index had a value of ca 0.7.Use of the Manhattan metric with catch standardized data indicated changes of community composition at the (more coastal) Bramble Bay sites with a roughly 3‐yr cycle and at the (more oceanic) Middle Banks sites with a roughly 51/2‐yr cycle.In the main analyses, sites were considered individually and in various groupings. It was shown that at Bramble Bay the optimum near‐annual periodicity gives a 50‐instead of a 52‐week year, and also that with increased numbers of sites in a grouping, the greater the significance of the near‐annual cycles.An hypothesis is outlined involving ‘counterbalancing’ between sites of depletion rates due to mobile predators or ‘disturbers’ (fish and prawns). The counterbalancing of recruitment rates is linked to depletion, and is in line with Caswell's (1978) somewhat theoretical paper in which he suggested the temporary action of predation opens up new cells for colonization.This hypothesis goes far towards ‘explaining’ the known patchiness of benthos in Moreton Bay. Also if mobile predators are a link between areas in proximity with different biotas the ecological unit (2 ‘commun