SCIENTIFIC ABSTRACT: We propose to observe an ensemble of atomic and ionic fine structure lines and the infrared continuum to characterize the interstellar medium and the global star formation process in a broad sample of normal galaxies. In this context, normal galaxies are those whose luminosity is dominated by stars and various amounts of ongoing star formation. Uniquely suited to the Infrared Space Observatory (ISO), these observations will probe the birth-sites of stars, yield the physical properties of the gas, dust and radiation field, and refine our picture of the global energetics in galaxies. We propose to study about 70 unresolved galaxies and selected regions of eight nearby, spatially resolved galaxies: NGC 6946, NGC 4449, IC 10, NGC 1313, NGC6744, NGC 2366 or NGC 6822, M101 or NGC 1365 and NGC 4258. We plan to use the LWS to probe the warm neutral phase with the [CII] line at 158 um, and [OI] at 63 um, and to probe HII regions with [NII] at 122 um, [NIII] at 57 um, and the [OIII] lines at 88 and 52 um. We propose to measure with PHOT the spectral energy distribution of our sources between 100 and 200 um to supplement IRAS data, and to obtain their images at 7 and 15 um using CAM. For the brighter sources, we will also obtain the 3-12 um spectrum using PHT-S. We plan then to study how these measured and derived physical parameters relate to the global properties of galaxies -- morphology and kinematics, environment, gas content, metallicity, and star formation efficiency. The ultimate goal of this program is to constrain the evolutionary processes in galaxies, discriminating for example between mechanisms for global star formation activation (cloud-cloud collisions or gravitational instability), and suppression (spatial dispersal or photodissociation and heating). Normal galaxies account for most of the luminous mass in the local Universe. The galaxies we would study sample wide ranges in total luminosity ($10^8-10^{10} L_{sun}$), infrared-to-blue luminosity ratio (0.3-50), morphology (S0 to Im; two or three Elliptical), and relative gas content. We exclude galaxies with active nuclei (Seyferts), but include a few objects with extreme infrared-to-blue ratio, or 60 um surface brightness. Spatial resolution in nearby galaxies allows us to study well-defined components such as HII complexes, spiral arms, bars and inter-arm disk regions, and variations in the properties of gas, dust and radiation within a disk. OBSERVATION SUMMARY: The proposed science objectives will be pursued by probing the ISM in star forming galaxies using various capabilities available on ISO. We use LWS for spectral lines, CAM for mapping the hot dust, and PHOT to detect long-wavelength emission from cold dust and for near-infrared spectroscopy. There are two groups of targets in this project: selected positions within ten nearby, resolved galaxies (proposal parts 6 and 7 of this project), and about 70 more distant, essentially unresolved galaxies (proposal parts 1 to 6). The planned observations are essentially the same for all targets except that some mapping is possible for the nearby sample. The data on the more distant objects will yield global fluxes, and rough sizes from CAM imaging. The basic set of observations consist of:\ LWS: A: [CII] 158 um, [NII] 122 um,\ B: [OI] 63 um, [OIII] 88 um,\ C: [NIII] 57 um, [OIII] 52 um;\ D: Full spectral scan for the few brightest sources;\ CAM: imaging at 6"/pixel in LW2 (7 um), LW3 (15 um) filters;\ PHT: photometry at 135 um and 200 um, using the C_200 module;\ resolved galaxies will be partially mapped at 60 um with C_100,\ and at 135 and 200 um with C_200;\ PHT-S: 3-12 um spectra will be obtained for the brighter targets.\ Wherever possible, especially for bright targets, all these data are collected for each object. However, various observations are eliminated for individual targets if they become too time-consuming. Thus PHT-S observations are only carried out for targets with an IRAS 12 um flux density greater than about 0.3 Jy. In the case of LWS spectroscopy, as the targets become fainter, the lines observed are reduced from sets A, B and C to A and B, and only to A for faint areas within nearby resolved galaxies. In the case of the very brightest sources, the LWS will be used to scan the whole spectrum. Some CAM and PHT observations for certain bright sources are not included because they already appear in the ISO Central Programme. Wherever CAM observations are planned, they include both LW2 and LW3 filters and use micro-scanning. Long-wavelength photometric observations are made with the PHOT-C instrument and complement the spectroscopy being carried out with the LWS.