Reef-scale, eddy-resolving numerical models are applied to discriminate between local trapping of neutrally buoyant passive material coming from a natal reef versus trapping of this material on reefs downstream. A hydrodynamic model is coupled with a Lagrangian (nongridded) dispersal simulation to map the movement of material such as passive larvae within and between natural reefs. To simplify the interpretation, a number of schematic reef shapes, sizes and spacings were devised to represent the most common cases typifying Australia's Great Barrier Reef. Prior investigations have shown that coral reefs on the Great Barrier Reef may retain material for times equivalent to the pelagic dispersal period of many species. This paper explores whether larvae are more likely to settle on the natal reef, settle downstream or fail to settle at all. The modelling neglects active larval behaviour and treats the vertically well-mixed case of notionally weightless particles only. The crown-of-thorns starfish larvae with a pelagic dispersal period of at least 10 days are one example of this case. Larvae are most likely to be found near the natal reef rather than its downstream neighbour, mostly because the currents take the vertically well-mixed material around, rather than onto, the downstream reef. Of all the simulations, the highest numbers were found on natal reefs (e.g. 8% after 10 days) while downstream numbers mostly varied between 0 and 1% after 10 days. Particle numbers equalised only when spacing between the two reefs was less than the reef length (6 km), or when the downstream reef was in the direct path of the larval stream.