Cytomegaloviruses (CMVs) colonize blood-borne myeloid cells. Murine CMV (MCMV) spreads from the lungs via infected CD11c+ cells, consistent with an important role for dendritic cells (DC). We show here that MCMV entering via the olfactory epithelium, a natural transmission portal, also spreads via infected DC. They reached lymph nodes, entered the blood via high endothelial venules, and then entered the salivary glands, driven by constitutive signaling of the viral M33 G protein-coupled receptor (GPCR). Intraperitoneal infection also delivered MCMV to the salivary glands via DC. However, it also seeded F4/80+ infected macrophages to the blood; they did not enter the salivary glands or require M33 for extravasation. Instead, they seeded infection to a range of other sites, including brown adipose tissue (BAT). Peritoneal cells infected ex vivo then adoptively transferred showed similar cell type-dependent differences in distribution, with abundant F4/80+ cells in BAT and CD11c+ cells in the salivary glands. BAT colonization by CMV-infected cells was insensitive to pertussis toxin inhibition of the GPCR signaling through Gi/o substrate, whereas salivary gland colonization was sensitive. Since salivary gland infection required both M33 and Gi/o-coupled signaling, whereas BAT infection required neither, these migrations were mechanistically distinct. MCMV spread from the lungs or nose depended on DC, controlled by M33. Infecting other monocyte populations resulted in unpredictable new infections.IMPORTANCE Cytomegaloviruses (CMVs) spread through the blood by infecting monocytes, and this can lead to disease. With murine CMV (MCMV) we can track infected myeloid cells and so understand how CMVs spread. Previous experiments have injected MCMV into the peritoneal cavity. MCMV normally enters mice via the olfactory epithelium. We show that olfactory infection spreads via dendritic cells, which MCMV directs to the salivary glands. Peritoneal infection similarly reached the salivary glands via dendritic cells. However, it also infected other monocyte types, and they spread infection to other tissues. Thus, infecting the "wrong" monocytes altered virus spread, with potential to cause disease. These results provide a basis for understanding how the monocyte types infected by human CMV might promote different infection outcomes.