Collection efficiency was measured for a wide range of surgical masks and certified respirator filters using a 0.55-microm latex sphere aerosol at 45 L/min. Results were used to select representative filters and surgical masks for subsequent biological aerosol challenges. Collection efficiency of 16 respirator filters and 5 surgical masks was determined using a Mycobacterium abscessus aerosol at 45 and 85 L/min and 30 and 70 percent relative humidity. The bioaerosol was measured using two instruments: (1) an aerodynamic particle sizer, which detects both biological and nonbiological aerosols, and (2) a modified six-stage Andersen sampler, which detects only those aerosols able to replicate. Filter penetration of the latex and biological aerosols was highest and most variable for the surgical masks. Latex aerosol penetration through respirator filters ranged from 0.3 to 10 percent for dust/mist (DM) filters and from 0.09 to 3.4 percent for dust/fume/mist (DFM) filters; high efficiency particulate air (HEPA) filter penetration was always less than 0.01 percent. Median penetration of the biological aerosol was 2, 0.4, and 0.02 percent for DM, DFM, and HEPA filters, respectively. Higher flow resulted in higher penetration for all filters (as expected) and changes in relative humidity exerted only minimal influence on the collection of bioaerosols. The two sampling instruments gave similar values of filter efficiency; the total particle counter was generally less variable than the viable particle sampler. If the instruments detect similar size ranges and the aerosol consists largely of viable organisms, it may be possible to employ a nonviable particle sampler to assess biological aerosol penetration through filters. Selecting a surrogate test organism requires careful evaluation and may depend on geometric size, shape, density, and biological state. For assessment of filter performance, a worst-case test should simulate the most penetrating particle size and effects of relative humidity on both filter media and biological organisms.