This paper is concerned with the dispersion of VIth stage nauplius larvae, free swimming cyprid larvae, crawling cyprid larvae and recently settled barnacles of the species Balanus crenatus, B. improvisus, B. amphitrite and Elminius modestus. It is shown (Chapter III) that numbers of VIth stage nauplii and cyprids of B. crenatus and B. improvisus and cyprids of Elminius modestus per unit volume show a tidal variation in the Western Wadden Sea. In addition to this there is a short period variation; it is thought that larvae do occur in groups in the water. The observations are concomitant with the assumption that cyprids are transported by tidal currents, that they sink to the bottom during periods of low current velocity, and that they are redispersed again in the water column when current increases. Mechanisms for group formation are discussed and it is postulated that groups are being formed by currents and accompanying turbulence. It is shown that numbers of larvae show a correlation with the amount of suspended matter in the water column. For this reason it is argued that retention of larvae in estuaries can be explained by a fully mechanical process, and that there is no need to suppose that the swimming behaviour of the larvae, induced or released by a tide-coupled environmental factor, influences retention. It is further shown that the field data from a number of authors, on oyster larvae, mussel larvae, barnacle larvae and medusae are not sufficient to sustain the conclusions that these animals further their retention in the estuaries by swimming. Attempts to characterize groups of cyprid larvae as to size were not successful. Laboratory experiments on the sinking, and redispersion of cyprid larvae by current have been described (Chapter IV). The majority of cyprids sink with a speed of 10 to 32 cm/min. Horizontal swlmming was hardly ever observed. Redispersion into the water column occurred at current velocities of 35 to 67 cm/sec. In view of the large tidal variations in the numbers of larvae in the water column it would be reasonable to expect variations in the settlement of cyprid larvae with the tide. Experiments to this end have been described (Chapter V) ; the results show that settlement occurs throughout the tidal cycle in low numbers, but a large proportion of total settlement takes place in 1 or 2 hours during early flood. During later flood, numbers of larvae in the water are higher, but greater current velocities prevent settlement. It is to be expected that the groupwise distribution of the cyprids in the water induces differences in settlement from place to place. Observations on such differences have been given (Chapter VI) : firstly on panels in a linear row, secondly on panels exposed in a regular 3 dimensional grid suspended from a raft. The linear series of panels showed that the numbers of settlers can be highly variable, even at distances as small as 5 m. This can be ascribed fully to the groupwise distribution of the larvae in the water. For panels on the raft differences in settlement can similarly be explained by transport of larvae by current. Publications of other authors, explaining differences in settlement in terms of light, current and swimming behaviour of larvae have been criticized, although it is shown that crawling behaviour of larvae may influence settlement density. Observations have been described on the pattern of settlement, especially the development of an even distribution on the substrate (Chapter VII). Even distributions develop only after a certain population density has been reached; the intensity of searching by the crawling larva and the spacing-out mechanism are of importance. It follows from the observations that these are different in Balanus amphitrite and B. crenatus; searching for previously settled individuals seems to be more intense in larvae of B. crenatus, although simulation experiments indicate (Chapter IX) that a large settlement during a short period may be of importance in this respect. It has been argued that natural mortality of young settled barnacles may have an influence on the settlement pattern; it is thought that density-dependent mortality could possibly change the pattern from uneven to even. Experiments to investigate this have not been very successful. Too heavy settlement developed unsuitable even patterns. A few factors, thought to be of importance in the development of settlement patterns, which are, however, difficult to investigate experimentally, such as intensity of searching of the crawling cyprids, age of cyprids, and numbers of settlers arriving per unit time, have been studied in a simulation model in a computer (Chapter IX). Some results of simulations influenced experimental work in the field, especially in the case of Balanus crenatus, where a high number of arrivals per unit time is necessary to develop an even distribution on the substrate. This was found in the field in accordance to earlier indications by simulation. Lastly, a number of applications of the results to the testing of antifouling paints has been given (Chapter X); these concern the number of replicates necessary for the testing, the patchiness of barnacles on aged antifouling paints, and a new test method based upon mortality rates of young settlers on aged antifouling paints as compared to those of non-toxic controls.