A powerful data processing methodology for analysis and classification of two-dimensional gels is introduced. The approach is based on correspondence analysis (CA) and ascendant hierarchical classification (AHC), and significantly differs from the more classical principal-component decomposition. Starting with a series of gels, each having a large number of spots, CA allows their representation in a factorial space of reduced dimension; classification into meaningful groups is then performed using AHC. Simultaneous representation of both spots and gels in the same space can be done. This precisely indicates the key spots pertinent for the classification, and therefore the characteristic proteins representative of a particular class of gels (i.e. of a particular disease or biological status). In addition, knowledge of these characteristic spots greatly simplifies the screening of future gels. After a brief overview of the Mélanie system for analyzing 2D gels, the theory of correspondence analysis and ascendant hierarchical classification is summarized. Equations are given that are easily ammenable to computation. How classification of two-dimensional gel electrophoretograms is accomplished is then detailed. Experimental results support the power of this approach.