SCIENTIFIC ABSTRACT ISO will provide for the first time the possibility to study interstellar dust over a large and continuous spectroscopic window (2.5 - 180 micron) with a resolution well suited for the study of solid state features arising from interstellar grains. By using the PHOT-S, and SWS instruments, we propose to take advantage of the full wavelength coverage to obtain a coherent set of informations on the chemical and physical nature of interstellar dust grains in molecular clouds and on PAH molecules in selected sources. Basically, three kinds of studies are proposed here: (i) spectroscopy of infrared sources embedded in molecular clouds with emphasis on volatile ices, (ii) low resolution spectroscopy of a statistically significant number of IRAS point sources located behind nearby molecular clouds, and (iii) spectroscopy of selected point like sources revealing some strong PAH features in the mid-infrared. The first two items refer to absorption spectroscopy and are mainly devoted to the problem of gas-grains interactions in the chemistry of molecular clouds. The third one is an emission spectroscopy study to provide, as far as possible, examples of complete spectral coverage of PAH emission in a few objects. This proposal is strongly coupled to laboratory experiments currently under development. OBSERVATION SUMMARY Most bands occuring in the solid state are broad and therefore a medium spectral resolution will be sufficient in most cases. Higher resolution will be required only for minor species and to separate blend of lines expected from complex mixtures to allow a finer interpretation of chemical process in interstellar grains. Also a very high signal-to-noise ratio is required in a few objects in order to measure potentially significant but minor species, which transitions occur at specific wavelengths.The strategy is: (i) to obtain complete SWS spectra from 2.5 to 45 micron (AOT SWS01), with a typical S/N of 30 except at the lower wavelengths, of a number of sources carefully selected on their IRAS-LRS spectra and other data when available. (ii) These sources need also to be observed with PHOT-S (AOT PHT40) because SWS will not be sensitive enough below 6 micron in these heavily obscured sources. This is why we request concatenation of PHOT-S and SWS observations of each source that we observe with the two instruments. (iii) For a subset of these sources we will also obtain grating scans (AOTSWS06) of the following ranges, given in micron, with the corresponding S/N: 2.38-2.82 (S/N>20), 3.3-3.35 (S/N>20), 3.9-4.6 (S/N>50), 5.3-5.6 (S/N=100), 5.6-6.0 (S/N=100), 6.3-6.5 (S/N=1000), 6.5-6.9 (S/N=100), 7.5-7.8 (S/N=100),9.4-9.8 (S/N=100), 14.5-16.5 (S/N=300). (iv) A similar strategy (AOT's SWS01 and PHT40) applies to a sample of 5 emission sources (PAHs). (v) Background sources toward nearby molecular clouds (Ophiucus, Taurus and Chamaeleon) will be measured with PHOT-S (AOT PHT40) in order to trace the variations in the dust properties across the cloud and in particular in regions far from the heating sources. Strong variations in the far-infrared emission have been evidenced by the IRAS data. A spectroscopic study would be very informative to address this problem. For this study it is necessary to obtain as many spectra as possible for a good sampling across the cloud; thus, we ask for concatenated observations of sources located in directions close to each other.