Bacteria load in saliva was experimentally manipulated, and the consequences for the measurement of salivary testosterone (T), dehydroepiandrosterone (DHEA), and cortisol (C) were examined. Healthy adults (n = 19) donated the first saliva sample upon rising after which they rinsed their mouths with water, waited 10 min, and donated a second sample. Samples were either left untreated or passed through a 0.22-microm filter and then frozen at -80 degrees C or incubated at room temperature (RT) for 10 days. Aliquots of each sample were cultured on agar to determine baseline and post-incubation (or freezing) bacteria load. Bacteria counts were not significantly influenced by rinsing (with water), were substantially reduced by filtration, and increased by incubation at RT. Average levels of salivary T and C, but not DHEA, were significantly lower in samples stored at RT than samples frozen the day of collection. The change in bacteria count induced by storing samples at RT was associated with a change in testosterone but not cortisol or DHEA. When samples were passed through a 0.22-microm filter bacteria counts were reduced, and the association between bacteria and testosterone was reduced to non-significant. These findings contribute to a growing body of literature revealing that the process of sample collection, storage, and handling can dramatically influence the accuracy of information generated when salivary biomarkers are integrated into research and clinical diagnostics.