Following the 1993 hantavirus pulmonary syndrome (HPS) epidemic in the southwestern United States, mammalogists and epidemiologists instituted long-term studies to monitor population density and prevalence of infection in rodents which constitute the reservoir for Sin Nombre virus (SNV). In this study, field techniques used in sampling small mammals for SNV infection were evaluated to determine if trapping and handling protocols were having significant effects on future trapability or mortality of animals. We compared rodent mark-recapture control plots, on which all rodents were simply measured, marked, and released on site, with experimental plots on which all animals were anesthetized with methoxyflurane, sampled for blood and saliva, measured, marked, and released. Blood samples were obtained from anesthetized animals on the experimental plots via a retro-orbital sinus puncture using a heparinized capillary tube. Dacron tipped oral swabs were used to collect buccal cells and saliva from the rodent's oral cavity. Field data were collected monthly from August 1994 to August 1996 at two sites in New Mexico (USA). Analyses were based on 3,661 captures of 1,513 individuals representing 21 species from three rodent families (Rodentia: Muridae, Heteromyidae, Sciuridae) and two species of rabbits (Lagomorpha: Leporidae). Overall, for most murid rodents (including five Peromyscus spp., Neotoma albigula, and Onychomys leucogaster) and one rabbit species (Sylvilagus floridanus), the handling/bleeding procedures had no significant effects on recapture rates or mortality. In contrast, several species of heteromyids (Dipodomys ordii and Perognathus flavus), one murid (Reithrodontomys megalotis) and one leporid (S. auduboni) suffered higher mortality rates, and heteromyid kangaroo rats (D. ordii and D. merriami) exhibited lower trapability as a result of the anesthesia and sampling procedures. In view of the overall non-significant influence of the sampling procedures on murid rodents, the anesthesia and blood/saliva sampling protocols described herein appear to be appropriate for hantavirus research, and may serve as a model for environmental monitoring of other zoonotic agents and their reservoirs.