I propose here an interesting historical scenario for galaxy formation that may find application to the problems of dwarf irregular galaxies in general and of I Zw18 in particular. My purpose is to present calculations of the C:N:O ratios within this simple model in an attempt to explain the apparent paradox, that characterizes I Zw18, of having very low concentrations of C, N and O in HII regions of a current starburst (some forty times smaller than solar) and yet having nearly solar ratios for C/O and N/O. I first calculate concentrations in a hot ambient medium that suffers a "Hubble-like flow" wind in a galaxy that has experienced several bursts of star formation. This hot matrix contains HI clouds whose collisions initiate the starbursts. The ambient gas which is found to have large and variable X\sbO, X\sbC and X\sbN concentrations, is mixed very slowly into the clouds, so that today a few percent of the cloud mass has been gathered from the hot medium by admixing, resulting in cloud concentrations comparable to those found in I Zw18. Different assumptions were tried for the details of the galactic history. The resulting ratios of the abundances in the clouds were found to be insensitive to these details.