We plan to investigate the relative importance of the warm and cool dust
emission, the dust opacity, and the characteristics of the dust-star
geometrical distribution in galaxies by studying multiwavelength spectral
energy distributions of a sample of star-forming galaxies. The ultraviolet and
optical spectra of our sample galaxies will be complemented by ISOPHOT
measurements of the dust emission at 60, 100, 135 and 200 um. Using the
spectral energy distribution of the dust emission in the range 40-240 um we
will investigate the cool dust emission, heated by the non-ionizing stars, and
how it relates to the warm dust emission, heated by the hot, young stars. From
the UV and optical spectra we can estimate the amount of ionizing and
non-ionizing energy absorbed by dust. This information will be used, in
conjunction with the ISO measurements, to study the energy balance between dust
absorption of UV and optical starlight and dust emission in the far infrared,
and to infer the dust opacity of the galaxies.

The long wavelength baseline of ISO is a requirement for the success of the
present investigation, since the cool dust emission peaks at wavelengths
greater than 100 um. The IRAS fluxes of most of our sample galaxies are known
with large uncertainties or are upper limits; ISO, thanks to its high
sensitivity, will provide accurate measurements for these objects.