The stellar content and stellar age distribution of the Pegasus Dwarf Irregular Galaxy are discussed in the light of deep CM diagrams for (V - R) and (V - I) color indices, complemented by Hα data. The CM diagrams of Pegasus show two structures-red tail and red tangle-which denote the presence of intermediate-age and old stars in the galaxy. After NGC 6822, Pegasus is the second dwarf irregular galaxy in which these structures are clearly detected, suggesting that the CM diagram of most dwarf irregulars could show these structures if photometry is deep enough. The youngest star in Pegasus is some 10 Myr old, but the galaxy shows little star forming activity over the last 100 Myr. However the CM diagram is populated by a large number of older stars which can be up to 10 Gyr old or more. These stars mainly concentrate in the red tangle. From the maximum extension to the red of the red tail, we estimate that the maximum metallicity of Pegasus' stars could be as large as Z=0.008. By means of a further development of the technique of artificial stars, we carried out a careful, deep study of the effects of crowding on the measured colors and magnitudes of stars. This is a necessary step, previous to the experiments with synthetic diagrams on which we are currently working to analyze the full star formation history of this galaxy. Data for 40000 artificial stars, with injected colors and magnitudes spread over a wide range, were added to the observed frames in several steps, and used to derive the most important crowding effects, namely the fraction of missing stars-or crowding factors-magnitude and color shifts, and external photometric errors, each as a function of magnitude and (V - R) and (V - I) color indices. The main effects detected from our tests can be summarized as follows: (l) Crowding factors, color and magnitude shifts, and external errors are a function not only of magnitude but also of color index. (2) In general, crowding produces a shift of blue stars to the red and of red stars to the blue. (3) There is no significant relation between external and internal errors; external errors are larger by a factor which varies from 1 to 7. Furthermore, we have tested a widely used, abbreviated alternative procedure which does not consider injected color indices, but only magnitudes of the artificial stars. It is much faster, but its results are inadequate for the synthetic diagrams tests. The weakness of the method lies in the fact that it overlooks the dependency of crowding effects on the color indices of stars. This dependency is large, especially for external errors.