at the time of sampling. Similar sampling at different times allows an estimation of patterns and rates of change in source strengths. Similar principles apply to the collection of liquid samples. If liquid reservoirs are well-stirred, a single sample can be representative of the entire reservoir.
Surface samples are prepared by swabbing or pressing a plate of culture medium or a sticky tape against a surface; such samples are useful for identifying obvious microbial contamination. Many samples must be taken to allow characterization of the surfaces in a space, and such data may or may not be relevant to allergen exposures.
Ideally, human respiratory exposure is measured using air samples taken near the breathing zone of individuals (personal sampling; Macher and First, 1984). Most allergen sampling, however, is done to characterize ambient aerosols. Air sample collection involves drawing a representative sample of the aerosol into a collection device and removing particles in an unbiased way and in a form that allows appropriate analysis.
Volumetric air samples for indoor allergen analysis are usually collected by suction devices. To accurately represent the aerosol, suction samples should be collected under isokinetic conditions in which ambient air flows into the sampler parallel to and at the same rate and direction of suction. In still air and in cases in which the sampling orifice is at an angle to ambient air movement, suction samplers tend to oversample small particles, which are easy to divert from their original path. When air is moving into the sampler orifice faster than the suction rate, small spores will tend to follow streamlines around the orifice and be undercollected. In most sampling protocols, isokinetic conditions are not present. For the small filter cassette samplers and the low flow rate suction impactors (e.g., the Andersen type) used in indoor environments, the error introduced is small. However, high-volume filter samplers and, possibly, the portable suction impactors (SAS, RCS) pull air into the sampler at a rate much higher than ambient air speeds; consequently, small particles are preferentially collected. It is also important to collect a small enough sample that the aerosol in the space is not changed during sampling. High-volume filtration devices process large amounts of air and can actually act as air cleaners.
Once particles have entered the sampler, they must be removed from the airstream. The two most commonly used methods are inertial impaction and filtration. Inertial impaction allows the collection of particles that are able to cross the airstream lines inside a sampler and thus stop at the collection surface. In general, large particles are more efficiently collected by inertial impaction than are small particles. The impaction samplers are often rated by the 50 percent cutpoint (the particle diameter at which 50 percent of particles entering the sampler will be retrieved; ACGIH, 1989). Ratings are set for the commonly used aeroallergen samplers listed in Table 6-1.