Small local dislocations in processed emulsions give rise to spurious scattering whose effect on tracks is indistinguishable from the multiple scattering produced by coulomb interaction. Because of this effect the atomic number of fast primary cosmic ray nuclei will be underestimated in more than 50% of the cases in experiments where multiple scattering of tracks and ionization measurements are employed for charge determination. The spurious scattering was investigated by measuring the track contours of 100 very long tracks due to energetic primary particles; the effect was present in comparable strength in all plates and all types of emulsions which have been investigated. It is too small to affect measurements on medium energy particle tracks (protons with energy below 600 MeV and heavy nuclei with energy below 300 MeV/ nucleon). It dominates, however, other sources or error and noise for tracks of energetic particles although it does not preclude occasional observations of very low scattering values. New methods for measuring various forms of noise have also been developed in the course of this work and the noise level for scattering measurements has been reduced below previously accepted values.
Spurious scattering is presumably largely responsible for the discrepancies which appear when one compares the primary charge and energy spectra derived from experiments involving scattering measurements with the corresponding spectra derived from numerous other experiments which employ different techniques.
If the experiments based on scattering measurements are omitted, the remaining evidence strongly favours a spectrum in which the energy per nucleon is nearly independent of atomic number for all primaries. It also favours a charge distribution which has a pronounced minimum for charges 3≤ Z≤ 5 and, therefore, yields a fairly low upper limit for the amount of interstellar matter traversed by primary cosmic ray nuclei.