We have been carrying out a large number of Monte Carlo simulations of extensive air showers, covering the energy range from 105 to 108 GeV, in an attempt to solve the problem of the primary composition in this energy range. In view of the results of the pbar-p collider at CERN, which confirmed our early conclusion that Feynman scaling is not valid in the central region at these energies, we have confined our calculations to a single model to handle the fundamental hadron-light-nucleus interactions, which is of the multiple-cluster type. The calculations have been carried out for proton and iron primaries. We present theoretical data, including longitudinal shower development curves, height of maximum distributions, energy spectra and correlation plots, together with compilations of experimental results. Comparison between new theoretical and experimental data leads to the conclusion that iron cannot be the dominating component in the primary radiation in the investigated energy range.