A pathogen infection can result in different outcomes, ranging from swift clearance of the pathogen to rapid death of the host. However, how such variation arises is not always well understood. This is particularly the case for bifurcating infection dynamics, which result in two distinct outcomes in the same host population: 1) terminal infections with rapid pathogen proliferation and imminent host death, and 2) persistent infections with lower, stable pathogen loads with no apparent effect on host survival. While there is theoretical agreement about how terminal infections arise, the emergence of persistent infections is currently debated. To shed light on this debate we experimentally investigated the hypothesis derived from a theoretical model, that pathogens persist in a specific body part within a tissue that protects the pathogen from the host immune response. For this purpose, we experimentally infected female Drosophila melanogaster with the bacterial pathogen Providencia burhodogranariea. Although the hypothesis was not supported by our data because the bacteria disseminated across the fly bodies, we observed spatial variation in pathogen clearance, which started most frequently in the abdomen. We also observed spatial variation in the persistent bacterial load, which was lower in the abdomen. These findings support the idea that the abdomen is less susceptible to infection or that bacteria survive less well in this location. Interestingly, our results additionally suggest that terminal infections are more likely to start due to bacterial proliferation in the thorax. We also extended a previous theoretical model to assess whether and how wounding related to thorax inoculation could facilitate the emergence of terminal infections. Taken together, our results suggest that instead of enabling persistent infections, protective niches for the pathogen might rather play a key role in the emergence of terminal infections.