Mesocosms filled with 80 liters of coastal seawater from Santa Monica, California, were used twice (June and November) to budget bacterial production and loss, as well as to assess the relative significance of viral lysis and protist grazing in bacterial mortality. Bacterial abundance was ~6 × 109 cells liter‒1 in June and 2 × 109 in November, with viral abundances ~2 × 1010 particles liter‒1 in June and 1.5 × 1010 in November. Incorporation of [3H]thymidine and leucine yielded essentially identical production estimates and allowed calculation of total bacterial mortality in these closed systems. Bacterial growth rates were 1–2 d‒1 in June and 1–3 d‒1 in November. Three independent lines of evidence indicated that bacterial mortality attributed to grazing by protists was about equal to that attributed to viruses: size fractionation of disappearance of labeled DNA, with a 50% reduction after protists were removed; comparison of protist grazing rates estimated with fluorescently labeled bacteria and virus production‐based bacterial lysis rates, with 40–50% of the total ascribed to viruses; and model‐based interpretation of the 3.3–4.6% of bacteria visibly infected with assembled intracellular viruses, suggesting that 24–66% of loss is due to infection. Redundant production and loss measurements as well as the independent loss process estimates agreed within ~30%, yielding a reasonably balanced budget. We believe the loss of bacteria to viruses reflects a significant dissipation of energy in this ecosystem and that viruses and protists contribute similarly to bacterial mortality.