The toxic haptophyte Prymnesium parvum regularly forms fish-killing blooms in inland brackish water bodies in the south-central USA. Along the Texas coast smaller blooms have occurred in isolated areas. There appears to be an increasing risk that harmful P. parvum blooms will propagate into open coastal waters with implementation of future water plans. These plans will include increased interbasin water transfers from the Brazos River, regularly impacted by P. parvum blooms, to the San Jacinto-Brazos Coastal Basin, which ultimately flows into Galveston Bay (GB). Persisting source populations of P. parvum in inland waters elevates this risk. Thus, there is a need for an increased understanding of how P. parvum might perform in coastal waters, such as those found in GB. Here, two in-field experiments were conducted to investigate the influence of various plankton size-fractions of GB water on inoculated P. parvum during fall and winter, periods when blooms are typically initiating and developing inland. Stationary- and log-growth phase P. parvum were used to represent high and low toxicity initial conditions. Results revealed that P. parvum could grow in GB waters and cause acute mortality to silverside minnows (Menidia beryllina). Depending on season and growth phase, however, P. parvum growth and toxicity varied in different size fractions. During the fall, P. parvum inoculated from stationary-, but not log-growth phase culture, was negatively affected by bacteria-sized particles. During the winter, bacteria and nanoplankton together had a negative effect on P. parvum inoculated from stationary- and, to a lesser degree, log-growth phase cultures. Intermediate- and large-sized grazers when combined with bacteria and nanoplankton had complex relationships with inoculated P. parvum, sometimes stimulating and sometimes suppressing population growth. Toxicity to fish occurred in almost all plankton size fractions. The inclusion of progressively larger sized plankton fractions resulted in trends of decreased toxicity in treatments inoculated with stationary-, but not log-growth phase P. parvum in the fall. In the winter, however, inclusion of larger sized plankton fractions resulted in trends of increased toxicity to fish in treatments inoculated with both stationary- and log-growth phase P. parvum. This study indicates that understanding P. parvum population dynamics in open waters of estuaries and bays will be challenging, as there appears to be complex relationships with naturally occurring components of the plankton. The observations that P. parvum is able to grow to high population density and produce fish-killing levels of toxins underscores the need for advanced risk assessment studies, especially in light of water use plans that will result in P. parvum invasions of greater size. (C) 2015 Elsevier B.V. All rights reserved.