We monitored Junin virus (JV) activity in rodent populations for 30 months at seven mark-recapture grids located in agricultural fields and adjacent roadsides and fence lines in endemic and nonendemic areas of Argentine hemorrhagic fever. Blood and oral swabs taken from rodents captured at five-week intervals were analyzed by enzyme-linked immunosorbent assay for JV antigen (Ag). Calomys laucha and C. musculinus were the most frequently captured rodents, making up 47% and 22% of captures, respectively. Of 41 Ag-positive captures, 37 were C. musculinus and four were C. laucha; 34 were from two trapping grids in the same locality. Antigen-positive Calomys were more frequently male (76%), and were found significantly more frequently among the oldest animals and the largest body mass classes. These patterns, combined with the greater mobility and higher frequencies of wounds among males than females, implicated horizontal transmission as the primary route of JV transmission between rodents. Seasonal maximum levels in JV prevalence (up to 25% of captured Ag-positive C. musculinus) occurred during periods of maximal population densities of Calomys. Spatial distribution of Ag-positive rodents reflected habitat preferences; most Ag-positive C. musculinus were captured from border habitats (roadsides and fence lines), and all Ag-positive C. laucha were captured in crop fields. These distinct, but previously undocumented, habitat preferences suggest that the disease in humans may be related to exposures to the primary reservoir species, C. musculinus, in border habitats rather than in crop fields.