The talk concerns a diagrammatic procedure for proof search in classical propositional logic P. The procedure may be extended to full predicative classical logic and may be (heavily) varied to accommodate intuitionistic logic and other logics. The idea is to start from a goal box (a box containing the formula one tries to derive) and to stepwise extend the diagram by `attaching´ diagrammatic elements (boxes which contain formulas and which are related by arrows) to bottom boxes of the present diagram. The logical elements analyse the formula of the bottom box, the premise elements, which provide from analysing a premise, justify the formula of the old bottom box in terms of a set of formulas (which occur in the new bottom boxes). I shall describe the elements as well as a couple of complications that are required to turn the method into a decision procedure for P. The procedure is complete in that, if applied to &#915