We present in detail the description and the analysis of two independent experiments using beams of ${}_{}{}^{16}{}_{}{}^{}\mathrm{O}$ and ${}_{}{}^{56}{}_{}{}^{}\mathrm{Fe}$. From their results it is concluded that the reaction mean free paths of relativistic projectile fragments, $3\le Z\le 26$, are shorter for a few centimeters after emission than at large distances, where they are compatible with values predicted from experiments on beam nuclei. The probability that this effect is due to a statistical fluctuation is <${10}^{-3\mathrm{}}$. The effect is enhanced in later generations of fragments, the correlation between successive generations suggesting a kind of "memory" for the anomaly. Various systematic and spurious effects as well as conventional explanations are discussed, mainly on the basis of direct experimental observations internal to our data, and found not to explain our results. The data can be interpreted by the relatively rare occurrence of anomalous fragments that interact with an unexpectedly large cross section. The statistical methods used in the analysis of the observations are fully described.

NUCLEAR REACTIONS ${}_{}{}^{16}{}_{}{}^{}\mathrm{O}$ and ${}_{}{}^{56}{}_{}{}^{}\mathrm{Fe}$ at $2A$ GeV on emulsion nuclei; reaction mean free paths of primary beams and projectile fragments measured and compared.

• Received 16 August 1982

DOI:https://doi.org/10.1103/PhysRevC.27.1489