QUANTITATIVE AND QUALITATIVE ANALYSIS OF INDUSTRIAL ODORS WITH HUMAN OBSERVERS *

The present paper reports a number of experiments on the response of the olfactory system to mixtures of industrial odors. The line of research is based on the assumption that man himself can be used as a measuring device for perceptual magnitudes as well as an analyzer of odor quality. Odor research on mixtures is a classical area of research, and the scanty literature reports instances of counteraction, independence, addition, masking, and synergism.lhs1fl Little effort, however, has been devoted to the perception of odor mixtures in general and to malodorous air pollutants in particular. Further, little interest seems to have been focused on possible mathematical descriptions of the process of odor interaction. One of tlie main aims of the experimental work to be presented has therefore heen to develop a mathematical model for perceptual odor interaction that is applicable to malodors. It is important that a model of odor interaction be used for predictive purposes in applied contexts. For example, such a model should predict the odor strength of the resultant effluent when process technology, for example, is changed. or when new filtering techniques are considered at the source. The methods adopted in the series of experiments to be reported here are various versions of the so-called direct scaling meth0ds.D. 20 The perceptual scales were calibrated so that the same unit of measurement was maintained throughout the various experiments. The research was carried out with the aid of an olfactometer described by Lindvall.17.23 Essentially, the olfactometer permits stable exposure conditions and rapid changes in concentration by means of various diluting systems. The observer is seated in a ventilated air-conditioned test chamber in whicli two odor-exposure hoods ;ire installed. The system permits natural hreathing. The problem of predicting the perceived odor strength of an n-odor mixture from the constituents was first reduced into a number of subproblems. In the first step, a mathematical model for odor interaction was developed by which the perceived odor strength of a mixture could be predicted from the perceived odor strength of the components of the mixture. In the second step, tlie psychophysical function for each component (substance) was determined. In the third step, these findings (model of interaction and psychophysical functions) were combined into predictions of the perceived odor strength of mixtures solely from the physical concentrations of the constituents. The research was divided into these three steps, rather tliaii conducted with use of a direct psychophysical approach, because of the uncertainty as to what is really the proximal stimulus12 for odor strength. Therefore, it seemed preferable to scale the odor strength of the single component odors as well as the mixtures and to