Understanding how pathogens are transmitted is central to the study of disease ecology. Despite the importance of host contact in disease transmission, there is limited empirical data in this area due to inherent difficulties of monitoring free-ranging vertebrates and the rarity of contact events in nature.  We investigated the role of contacts in a directly transmitted host-pathogen system (deer mice-Sin Nombre virus (SNV)) by developing and testing a novel surveillance system for use under natural conditions. In October 2007, we trapped and marked deer mice at a field site where deer mouse densities and SNV prevalence have been documented since 2002.  Prior to surveillance, animals were live trapped for three nights using a web configuration of 148 traps. All trapped animals were PIT tagged and deer mice were tested for SNV antibodies with an IgG ELISA. Following trapping, the surveillance system, which consists of nine surveillance stations spaced 50m apart in a 3x3 grid formation, was operated for three nights.  Each station consisted of an IR CCTV camera (model SCR351-HN1), a Biomark® data logger with PIT tag antenna, and a foraging tray containing 3g millet in 1l of sand. Antennas under each foraging tray recorded PIT tags within range of 0.66m.  Cameras mounted 0.66m above each foraging tray provided an observational area of ~0.78m2.  We defined “contacts” as the presence of two or more  animals within the observational area. Data from each station were transmitted to a central computer and integrated using software (Time Science). 
Results/Conclusions

We trapped and marked 6 deer mice in 444 trap nights.  This density (1.9 deer mice/ha.) was the lowest ever recorded over the 6 previous years of sampling at this site.  All six marked deer mice were recorded 143 times in a single night.  Contacts were recorded for five of the six marked animals for a total of ten contacts.  Three of these contacts consisted of several seconds of direct physical contact, the type responsible for SNV transmission.  Due the absence of SNV infection in the six marked deer mice we were unable to use host contacts to model transmission. We will continue surveillance of this population monthly beginning in March of 2008. The work to date demonstrates that deer mouse contact behavior can be observed in free-ranging populations. Further, the differences in the type of host contact observed in this study, physical versus non-physical, are an important addition to disease transmission models.