Marine Biology109, 157-166 (199l) Marine .............. BiOlOgy @ Springer-Verlag 1991 Vertical distribution of fish larvae in the Great Barrier Reef Lagoon, Australia J.M. Leis Division of Vertebrate Zoology,The Australian Museum, P.O. Box A285, SydneySouth, New South Wales 2000, Australia Date of final manuscript acceptance: January 11, 1991. Communicatedby G.F. Humphrey,Sydney Abstract. The vertical distribution of shorefish - primar- ily reef fish larvae in the upper 20 m in relatively shal- low (< 30 m) waters of the Great Barrier Reef Lagoon near Lizard Island was investigated from 27 February to 7 March 1983 during both day and night. Four strata were sampled by neuston net and opening/closing bongo net: neuston (0 to 0.1 m), upper (0 to 6 m), middle (6 to 13 m), and deep (13 to 20 m). Taxon-specific patterns of vertical distribution which changed little ontogenetically, were found for the 50 taxa (in 24 families) analysed. Ver- tical distribution was highly structured during the day, and with few exceptions was nearly unstructured at night. Most taxa had highest concentrations deep in the water column during the day, but in any given stratum some taxa had highest concentrations. Day/night changes in pattern apparently were due to randomization or spread, rather than active migration. Related taxa had similar patterns. Similarity analysis including 211 taxa produced three groupings of samples: day neuston; day upper/mid- dle; and day deep plus all night samples. The sampling strategy of using drogues to tag water parcels for subse- quent resampling was compared with one of sampling at a fixed point. The drogue strategy was not superior to the fixed-point strategy as measured by dissimilarity indices, but this may differ among strata. Introduction Fish larvae live in a three-dimensional medium, and while their horizontal distributions have been widely studied, vertical distributions have received less attention, partic- ularly in tropical waters (Leis in press). However, an un- derstanding of vertical distribution is essential for pro- gress in many aspects of larval fish ecology. Unless verti- cal distribution and its temporal variations are known, sampling cannot be planned properly, horizontal distri- butions cannot be fully interpreted, the relations between larval fish concentrations and variables ranging from temperature to food availability cannot be determined, and movement of larvae by currents cannot be under- stood or predicted. For example, the vertical distribution of herring larvae profoundly affects their transport by currents, as has been shown by the use of a realistic, three-dimensional model of larval dispersal (Bartsch 1988). Because the vertical distributions of reef-fish larvae are not well known, recent models have unrealistically assumed either uniform vertical distribution (e.g. Williams et al. 1984), or some arbitrary vertical distribu- tion pattern (e.g. Doherty et al. 1985). Further, vertical distributions of many types of fish larvae change dielly or with age, and this complicates any attempt to study verti- cal distribution or to apply the results of such studies (Pearre 1979). Study of vertical distribution is study of larval be- haviour, albeit indirectly and on a population, rather than an individual, level. Taxon-specific, age-specific, and perhaps location-specific patterns are to be expected. Therefore, it is all the more important that vertical distri- butions actually be determined rather than assumed or extrapolated. Little is known of the vertical distribution of tropical shorefish larvae. Most of the few available studies (Aboussouan 1965, Dekhnik et al. 1966, Watson 1974, Leis 1978, 1986, Liew 1983, Robison 1985; all reviewed by Leis in press) are not published in the primary litera- ture, thereby limiting accessibility, and they are uneven in quality. However, these studies agree that there are tax- on-specific patterns of vertical distribution which seem to have few spatial or temporal variations, except day/night differences. Few taxa were examined in more than one study, so it is unclear if the patterns are similar for related taxa. Relatively little attention has been given to ontoge- netic changes in vertical distribution. The goal of the present study was to determine the vertical distribution on shorefish (primarily reef fish) lar- vae both day and night in relatively shallow (<30 m), sheltered waters of the Great Barrier Reef Lagoon over a period of several days. Given the high species diversity of the tropics and the expectation of taxon-specific vertical distributions, the larvae were identified to the lowest level