Bacteria undergo cycles of growth and starvation, to which they must adapt swiftly. One important strategy for adjusting growth rates relies on ribosomal levels. While high ribosomal levels are required for fast growth, their dynamics during starvation remain unclear. Here, we analyzed ribosomal RNA (rRNA) content of individualSalmonellacells using Fluorescence In-Situ Hybridization (rRNA-FISH). During the transition from exponential to stationary phase we measured a dramatic decrease in rRNA numbers only in a subpopulation, resulting in a bimodal distribution of cells with high and low rRNA content. We showed that the two subpopulations are phenotypically distinct when subjected to nutritional upshifts. Using a transposon screen coupled with rRNA-FISH, we identified two mutants acting on rRNA transcription shutdown and degradation, that abolished the formation of the subpopulation with low rRNA content. Our work suggests thatSalmonellaemploys a bet-hedging strategy in regulating ribosomal levels that may be beneficial for survival.