Description: Highlights: • Belt Sea cod, plaice and flounder differentiate in their specific egg densities. • Ontogenetic egg density increase in stage IV cod eggs elevates modeled egg mortality. • Drift model indicates retention in western Baltic in cod and flatfish yolk sac larvae. • No eastward transport to Arkona Sea or Bornholm Sea until end of yolk sac stage. • Ambiguity in flounder egg density could reflect more complex population structure. Abstract: Vertical distribution is an important feature of pelagic fish eggs and yolk sac larvae impacting their survival and dispersal, especially in heterogeneous and highly variable estuarine environments like the Baltic Sea. Egg densities determining the vertical distribution pattern were experimentally ascertained for cod (Gadus morhua), plaice (Pleuronectes platessa) and flounder (Platichthys flesus) from the western Baltic Sea. Plaice eggs floated at lower mean (±standard deviation) density range (1.0136 ± 0.0007 g cm−3) compared to cod (1.0146 ± 0.0009 g cm−3) and flounder eggs (1.0160 ± 0.0015 g cm−3), which floated on the highest density level. In flounder egg diameter was significantly related to egg density and in cod a weak correlation could be found between egg dry weight and density. All other relationships between female size, egg size, egg dry weight and egg density were not significant for any of the species. Available egg density data for Baltic Sea cod, plaice and flounder are summarized considering ICES subdivisions and stock management units. A hydrodynamic drift modeling approach was applied releasing drifters in the Belt Sea continuously from December to May, covering the species’ spawning seasons. The model implemented experimentally derived egg density ranges and included ontogenetic egg density modifications for cod eggs, increasing egg density from a late egg development stage to first hatch. A drifter was removed from the model, i.e. considered dead, when its initially prescribed density value exceeded the density range available at the temporally resolved geographical positions along the drift trajectories. Highest survival occurred during releases in April and May but no cohorts survived if they were drifted east into the central Arkona Basin or the central Baltic Sea, irrespective of whether a major Baltic Inflow (1992/1993) or a stagnation-year (1987/1988) was simulated. The dispersal characteristics of the surviving yolk sac larvae of all three species reflected retention within the Belt Sea or northwards transport through the Great Belt into the Kattegat and partly into the Skagerrak. There was no successful transport to more eastern Baltic areas past a hypothetical line from the island of Moen (Denmark) to Kap Arkona on Rügen Island (Germany).