Viruses infecting bacteria (phages) are the most abundant and ubiquitous entities on Earth and likely critical to any ecosystem, as they influence nutrient cycling, mortality and evolution. Ultimately, their impact depends on whether phage—host interactions lead to intracellular phage coexistence (temperate phage) or cell death (lytic phage). Temperate phages in the lysogenic cycle replicate their genome (either integrated into the host chromosome or extrachromosomally), until induced to become lytic, when they create and release progeny via cell lysis. While knowledge on lytic versus lysogenic outcomes is vast, it largely derives from few model systems that underrepresent natural diversity. Further, less is known about the efficiency of phage—host interactions and the regulation of optimal versus sub-optimal lytic infections, which are predicted as relevant under environmental (nutrients, temperature) and host (availability, density) conditions that are common in the ocean. In this dissertation I characterize the phage—host interactions in a new marine model system, phage ϕ38:1 and its Cellulophaga baltica bacterial host, member of the ubiquitous Bacteroidetes phylum. First, I show ϕ38:1’s ability to infect numerous, genetically similar strains of the C. baltica species, two of which display contrasting infection outcomes–lytic versus sub-optimally lytic or lysogenic on the original versus alternative hosts, respectively. Second, I collaboratively apply new gene marker-based approaches (phageFISH and geneELISA) to study ϕ38:1’s infection at the single-cell level and show that it is sub-optimal on the alternative host, rather than lysogenic. Third, I collaboratively develop whole-genome transcriptome datasets for ϕ38:1 infecting both, the optimal and sub-optimal hosts, to characterize the cellular response to infection and hypothesize potential transcriptional and post-transcriptional regulation of the sub-optimal infection. Together, these findings advance our knowledge of naturally-occurring phage—host interactions with a focus on nearly-unstudied sub-optimal infections.