This study has shown that Alexandrium catenella and Protoceratium reticulatum are infected with different species of bacteria. Using 5-cyano-2,3-di-4-tolyl-tetrazolium chloride (CTC), a substrate that becomes fluorescent after being reduced by electrons produced during bacterial respiration and combined with confocal microscopy analysis, live intracellular bacteria were observed. It was possible to determine that these dinoflagellates had a heavier bacterial load when the culture was in stationary phase as compared to the log phase. We have termed this observation bloom inside the bloom . Whether the high number of intracellular bacteria observed is a result of intracellular bacteria multiplication or re-infection from broken cells with heavy bacterial infection is unknown. Additionally, the bacteria isolates from each dinoflagellate are capable to re-infect both dinoflagellates regardless of their origin. Furthermore, when sodium nitrate and sodium phosphate are supplemented into the culture medium the bacterial multiplication pattern in both dinoflagellates is modified. Finally, based on this study, we propose the hypothesis that the interaction between intracellular bacteria and dinoflagellate is bimodal: at the beginning of the growth curve is mutualistic and at later stages, bacteria become parasitic, killing the host cell. This hypothesis could explain the suddenly disappearance of blooms in nature, without excluding other factors