58 CHEMICAL SENSORS. VOLUME 6: CHEMICAL SENSORS APPLICATIONS The reason for the very large number of odors that can be detected is that although the specificity of each receptor class may be low, the combination of several specificity classes yields a very large information content. Mimicking this information handling has led to new strategies in measurement technology: Instead of measuring specific parameters, it has become advantageous to determine quality properties. This is done by collecting as much information as possible concerning the specific measurement situa- tion. The information is then treated with multivariate data analysis (MVDA), which is a very powerful tool to extract useful information from a large data set. Examples include speech recognition (in analogy with the auditory sense), vibration analysis (in analogy with the sense of feeling), and image analysis (in analogy with the sense of vision). A commercial system for quality assessment of grain, The Grain Analyzer by FOSS, is based on MVDA analyses of NIR spectra. 1.2. THE CONCEPT OF ELECTRONIC TONGUES The development of biomimetic sensing systems has been inspired by nature. These systems are based on arrays of sensors with partially overlapping selectivities, and the data obtained are treated with MVDA. The first concept of this kind was described in 1982 by Persaud and Dodds (1982), where an array of gas sensors was subjected to various gases or odors and the resulting response patterns were treated with multivariate data analysis. This system simulated the olfactory system and could thus be regarded as an