In order to investigate a possible difference in the mean free paths for shower production by protons and by $\pi$ mesons of very high energy, an experiment was carried out at Echo Lake, Colorado (altitude 3260 m) in which the rate of production of penetrating showers by the atmospheric $N$ component was studied as a function of the absorber thickness up to a total of 650-g/${\mathrm{cm}}^2$ Pb, or about four proton mean free paths. A difference in the cross sections would then result in the "filtering out" of the component with the shorter mean free path, and manifest itself in an attenuation curve which is the superposition of two exponentials rather than a single exponential curve. By splitting the top absorber into three sections and using three top trays instead of the usual one, with the rigid condition that only a single counter in each tray be discharged, losses due to the complete absorption of all secondaries of an interaction occurring in the top absorber were minimized and could be estimated from the results of the runs with smaller absorber thicknesses in which the lead absorber was distributed in various ways between and above the trays. Three further counter trays, placed in a lead pile, were used to select five types of showers of different multiplicity. The average energy of the primaries of these events were estimated to be about 5.3 Bev ("low-energy group"), 8.3 Bev ("medium-energy group"), and 16-20 Bev ("high-energy group"). Even after correcting for spurious events, the rates of the two latter groups—for which a substantial $\pi$-meson component can be expected—do not fit a straight line on a semilogarithmic plot, thus indicating that the mean free paths of the two components are not identical. Under the assumption of equal detection probability for the interactions of both types of primaries, the mean free path for shower production by $\pi$ mesons of both energy groups is found to exceed 320 g/${\mathrm{cm}}^2$. It must be stressed that these results should not be interpreted as a proof for an interaction cross section smaller than the geometrical nuclear cross section, but rather as an indication that in collisions of very energetic $\pi$ mesons, the number of $\pi$-meson secondaries emitted differs from that resulting from proton collisions, perhaps because of a larger probability for the production of other secondary particles.

• Received 7 June 1954

DOI:https://doi.org/10.1103/PhysRev.97.151