The end of the 1970 decade has provided exciting new results in cosmic ray astrophysics both in theory and observations. To mention only two of the many results: (i) the discovery of large isotopic anomalies in cosmic ray neon and possibly Mg and Si compositions as compared to the solar system matter (Fisher et al., 1976; Garcia-Munoz et al., 1979; Mewalt et al., 1980; Wiedenbeck et al., 1981) has provided for the first time a direct proof that a specific component of cosmic rays is accelerated near active sites of nucleosynthesis; (ii) the great advances made in understanding shock wave acceleration have given a great impetus to the problem of the origin of cosmic rays in allowing a direct comparison between theory and observations in a wide variety of physical conditions (Axford 1981). Yet as exciting and astrophysically rewarding as these results may be, they fall short of achieving the full potential of cosmic ray astrophysics in the study of sources and propagation of cosmic rays in the Galaxy. This information can best be obtained at relativistic energies, where composition changes due to solar modulation and ionization losses are negligible and where secondary cosmic ray production in the interstellar medium is minimized, with nuclear interaction cross sections being constant or varying only slowly with energy. However at relativistic energies, where the results are simpler to interpret, the mass measurements are more difficult to perform. In order to achieve these aims a radical improvement in satellite instrumentation was necessary, namely the transition from the small solid state telescope working beautifully in the lower energy range (~ 100 MeV/n) to the very large Cerenkov telescope providing high resolution and high statistics in the relativistic energy range.