According toNikolskijet al. (1957), the number of nuclear-active (\({\mathcal{N}}\)) particles in air showers increases fairly slowly, asE0.2, for showers ranging in energy from 1013 eV to 1015 eV. At energies >1015 eV the increase is more rapid and is ∝E. This variation in the\({\mathcal{N}}\)-component with the size of the shower and certain other observed features of air showers have been interpreted previously as an indication of a change in the nature of nuclear interactions at very high energies or as an indication of specific details in the nature and composition of the primary cosmic radiation, such as the existence of a cut-off in the magnetic rigidity of the particles and the predominance of heavy primaries at very high energies. In this paper an attempt is made to explain the above observed features in terms of certain characteristics of nucleon-nucleus (\(n - {\mathcal{N}}\)) pion-nucleus (\(\pi - {\mathcal{N}}\)) collisions; the term « nucleus » here refers to an « air nucleus ». The main aspect of the (Nikolskij) curve, for the abundance of\({\mathcal{N}}\)-particlesvs. shower energy, considered here, is not the rapid increase of the\({\mathcal{N}}\)-component in showers of energy >1015 eV, nor the existence of a change in the abundance of\({\mathcal{N}}\)-particles in going from small showers to big showers, but rather the fairly slow increase in the number of\({\mathcal{N}}\)-particles in showers of energy (1013÷1015) eV. This slow increase, which is contrary to general experience in cascade calculations, puts stringent limitations on the choice of parameters that characterize\(n - {\mathcal{N}}\) and\(\pi - {\mathcal{N}}\) collisions. It is shown that\(n - {\mathcal{N}}\) and\(\pi - {\mathcal{N}}\) collisions are largely elastic and that the number of secondary nucleons produced in such collisions is quite small (of the order of one or two); further, in a\(\pi - {\mathcal{N}}\) collision, the energy going into the nucleon component is a very small fraction of the energy of the incident pion. These features also result in a rapid increase in the\({\mathcal{N}}\)-component of big showers; this arises as a result of the complicated role played by charged pions owing to their finite life time. In small showers the charged pions have only small energies and hence they mostly decay and do not contribute appreciably to the\({\mathcal{N}}\)-component. In big showers, however, the pions have sufficiently high energy to make nuclear interactions before decaying and owing to the large multiplicity of pions produced in\(n - {\mathcal{N}}\) and\(\pi - {\mathcal{N}}\) collisions, there is a rapid increase in the number of\({\mathcal{N}}\)-particles with increasing energy of the shower. Some experiments are suggested to distinguish between the various possible explanations.