1: \begin{abstract}

2: We show that the proof of Luna's conjecture about the classification of general wonderful $G$-varieties can be reduced to the analysis of finitely many families of primitive cases. We work out all primitive cases arising with any classical group $G$.

3: \end{abstract}

4: