The reported growth rates of Karenia brevis vary from 0.2 to 0.5 divisions per day, both in laboratory and field populations observed. This growth rate alone is not sufficiently high to account for its dominance in the water column. However, careful studies have not been carried out to determine if the documented slow growth rate of K. brevis persists throughout bloom development, or if different rates of growth might characterize bloom initiation, growth, maintenance, and termination phases. For example, it is conceivable that an "explosive" growth stage occurs early during bloom initiation/growth stages that serves to boost the population size, which has not been previously documented. Like most autotrophic dinoflagellates studied, cell division in K. brevis blooms is phased to the diel cycle. Diel phasing of cell division imposes a maximum potential growth rate of 1 division per day in dinoflagellates. The occurrence of an "explosive growth stage" would require the release of K. brevis cells from mechanisms which regulate this circadian rhythm. To address this question, diel phasing of the cell cycle of K. brevis was first documented in laboratory isolates of K. brevis. In situ diel studies were then carried out on naturally occurring blooms of K. brevis during the R/V Pelican cruise (1996) and during ECOHAB process cruises carried out during years 1-5 of ECOHAB Florida (1997-2001). Cell cycle distribution of K. brevis cells and correlation of cell cycle events with vertical migration was determined by sampling at multiple depths (minimum of surface and bottom, depending on depth). For shipboard field studies, whole water samples of K. brevis (1-2 L) were collected every 2 or 3 h throughout a diel cycle, fixed in glutaraldehyde and stored in the dark. Further processing of samples was carried out in the laboratory following completion of the cruise. The flow cytometry method method of Van Dolah and Leighfield was used (1999. J. Phycology 35:1404-1411). Glutaraldehyde-fixed whole water samples were filtered through a 100 micrometer nitex screen to remove large phytoplankton/zooplankton/debris, then filtered by gravity flow through a 10 micrometer nylon screen to collect K. brevis cells. Cells were collected from the screen by rinsing with 2% glutaraldehyde into a 50 ml polypropylene centrifuge tube, then centrifuged at 1000 x g. The cell pellet was then treated with 2 ml of 20 degrees C methanol overnight to remove pigments. Methanol extracted cells were collected by centrifugation (500xg for 3 min) and resuspended in phosphate buffered saline + 0.5% Tween 20 containing 10 micrograms. mL-1 propidium iodide (PI, Sigma, St. Louis, MO) and 10 mg/ml Rnase (Sigma, St. Louis, MO). DNA analysis was carried out on an Epics MXL4 flow cytometer (Coulter, Miami, FL) using a 5 W argon laser with a 488 nm excitation wavelength and 635 nm emission wavelength. Cell aggregates were eliminated by gating all histograms within the linear range using a peak-area cytogram for PI fluorescence. Cell cycle distribution was analyzed using Multicycle software (Phoenix Flow Systems, San Diego, CA).