We propose to use ISOPHT-S, ISOPHT-P and the SWS grating to obtain low resolution spectra and photometry from 2.5 to 25 um of a sample of 50 IRAS galaxies. We have two primary scientific goals. First, we will use the PAH and silicate dust bands and available spectral lines to discern the origins of the large IR luminosities of these galaxies: starbursts or active galactic nuclei(AGNs). Strong PAH emission is indicative of a starburst nucleus, while the mid-IR line ratios derived from the species [ArII], [ArIII], [SIV], and [NeII] reflect the earliest stellar type still on the main sequence, hence the age of the starburst. Lack of PAH emission, together with emission from more highly ionized species like [ArV], [ArVI], [MgIV], and [MgVIII], signal the presence of an AGN. Lines from the refractories (e.g. [NiII]), trace shocks likely associated with supernovae, or galaxy-wide collisions. Second, our proposed 2.5 to 25 um spectrophotometry is essential to validate the use of far-IR colors as luminosity indicators in the deep cosmological surveys proposed both in the ISO core program (6.75 and 15 um surveys) and in the proposed WIRE mission (12 and 25 um surveys). Nearly all the sources detected in these two cosmological surveys will have intrinsic far-IR luminosities in excess of 10^10 L(solar). Therefore, our two scientific goals (discerning AGNs from starbursts, and providing calibration for deep surveys) are complementary: they both require a statistically significant sample of IR luminous galaxies at modest redshifts. We have selected 50 galaxies from the IRAS Bright Galaxy Survey with far-IR luminosities at and above 1.5 x 10^10 L(solar). The proposed observations will therefore investigate both the origins of the IR luminosity and provide a framework for understanding the evolution of IR-luminous galaxies in the early Universe.