The investigation of the excitation spectrum of baryons provides important information on many open questions in baryon spectroscopy. The key to any progress is the identification of the effective degrees of freedom leading to a qualitative understanding of strong QCD. The problem of missing resonances predicted by quark models is discussed on the basis of recent experimental results of the CB-ELSA experiment at the e− accelerator ELSA in Bonn. First hints for resonance production are given and cascades of the type \({N^{ * * }}\left( {{\Delta ^{ * * }}} \right) \to {N^ * }\left( {{\Delta ^ * }} \right) \to p{\pi ^0}{\pi ^0}\left( {p{\pi ^0}\eta } \right) \) are observed. Indications for a Δ-resonance around 1900 MeV are seen. The latter is particularly interesting if it had negative parity because a confirmation of this state would be in contradiction with almost all constituent quark models, (S. Capstick, N. Isgur, Phys. Rev. D 34, 2809 (1986); U. Lohring et al., Eur. Phys. J. A 10, 309 (2001)) for instance. Both, the quark models using one-gluon exchange and the quark model using instanton-induced forces as short-range residual quark-quark interaction predict the three states Δ 5/2− (1930), Δ 3/2− (1940) and Δ 1/2− (1900)at masses in the 2200 MeV region.