Gamma rays of energy in the range 30 MeV-several GeV, observed by the satellites SAS-2 and COS-B, are emitted in the interstellar medium as a result of interactions with gas of cosmic-ray nuclei in the GeV range (π° decay γ rays) and cosmic-ray electrons of energy > 30 MeV (bremsstrahlung γ rays). W. Hermsen has presented at this conference the γ ray maps of the Galaxy in three “colours” constructed by the COS-B collaboration; the information in such maps is supplemented by radio-continuum studies (see lecture by R. Beck), and is a useful tool for studying the distribution of gas, cosmic rays (c.r.) and magnetic fields in the Galaxy. The variables in this problem are many:large-scale (~ 1 kpc) and small-scale (~10 pc) distributions of c.r. nuclei, of c.r. electrons, of atomic and molecular hydrogen, of magnetic fields, fraction of the observed radiation due to localized sources, etc. Of these, only the distribution - or at least the column densities - of atomic hydrogen are determined in a reliable way. Estimates of the amount of molecular hydrogen can be derived from CO observations or from galaxy counts. The radio and gamma-ray data are not sufficient to disentangle all the other variables in a unique fashion, unless a number of assumptions are made (e.g. Paul et al. 1976). Still, the COS-B team has been able to show that :a) there is a correlation between the gamma-ray emission from local regions, as observed at intermediate latitudes, and the total column density of dust, as measured by galaxy counts. The simplest interpretation is that the density of c.r. nuclei and electrons is uniform within 500 pc of the sun, and that dust and gas are well mixed. Then, γ rays can be used as excellent tracers of local gas complexes (Lebrun et al. 1982, Strong et al. 1982).b) In the same way, the simplest interpretation of the γ-ray emission at energy > 300 MeV from the inner Galaxy, is that c.r. nuclei and electrons are distributed uniformly as well : there is no need for an enhanced density of c.r. in the 3–6 kpc ring; on the contrary, even assuming a uniform density of c.r., the γ-ray data are in conflict with the highest estimates of molecular hydrogen in the radio-astronomy literature (Mayer-Hasselwander et al. 1982).c) In the outer Galaxy, the gradient of c.r. which had become apparent in the early SAS-2 data can now, with COS-B data, be studied in three energy ranges. A gradient in the c.r. distribution is only required to explain the low-energy radiation, which is dominated by bremsstrahlung from relativistic electrons (Bloemen et al., in preparation).