The strict requirements of similarity theory are shown to be so numerous and restrictive that complete modeling of combustion processes is practically impossible; all successful modeling so far has involved deliberately ignoring many of the similarity rules. The paper reviews some of the more notable examples of this partial modeling, and mentions the physical facts which underlie their success. It is emphasized that every combustion-modeling technique represents a compromise between the requirements of cheapness and reliability.

The use of a secondary jet (usually a fuel jet) for flame stabilization is discussed and the fuel penetration concept used for correlating data for both subsonic and supersonic main flows. The fuel penetration height has been shown to be the characteristic dimension for a jet flame stabilizer for both the subsonic and supersonic cases.