Light N=Z nuclei show pronounced clustering for exited states close to the decay thresholds. This fact is expressed in the Ikeda-Diagram. In neutron-rich nuclei, low lying states close to the neutron drip-line show also pronounced clustering with α-particles and heavier closed shell clusters. For the valence neutrons a concept based on molecular orbitals can be used for their single particle wave functions. Covalent binding between α-particles due to valence neutrons and a related shallow local potential is now known to be responsible for particular structures in light neutron rich nuclei, like dimers in Be-isotopes, and trimers in C-isotopes. A local potential as between two α-particles can be defined between other clusters, for example between α-particles and 16 O, giving rise to intrinsic reflection asymmetric molecular structures in 21 Ne as parity doublets. An Extended Ikeda Diagram can be constructed based on these concepts with clusters and covalent neutrons. §1. IntroductionWe can look back into 30 years of studies of clustering phenomena in nuclear physics. There has been a steady evolution of the field, with a particularly strong push in recent years due to the study of neutron rich exotic nuclei. The earlier work was focused on light nuclei with N=Z and their reactions, which has been on the forefront of research, in particular due to the observation of the heavy-ion resonances. Many theoretical concepts have been clarified in the last decades, specifically the equivalence of the shell model description with the cluster model approaches, once antisymmetrisation is considered in its full consequence (see also Ref. 1)).The shell model approaches made big progress because they were able to include larger and larger model space. However, the need for cluster models can be illustrated by the fact that the most extensive shell model calculations 2) do not produce the famous second 0 + -state of 12 C. The long standing problem how to define a cluster has been a line of thought all this time. 1) One approach is perfectly illustrated by the diagram 3), 4) proposed by Ikeda in 1968. In Fig. 1 this diagram is shown. The concept suggests that states of a particular cluster-(sub)structure appear close to the threshold for the decay of the composite system into these "fragments". This can actually be observed with rising excitation energy starting from the bound states of normal nuclei, as well as function of decreasing binding, when approaching the dripline. The main physical effect of the diagram resides in the properties of the nucleon-nucleon interaction, which gives rise to saturation and the stabilisation of clusters with particular "magic" numbers, and their mutually reduced interaction.