5 Conclusions

Observations with high spatial resolution are necessary to assess the complicated structure of high-mass star-forming regions. Adaptive optics imaging at near-infrared wavelengths provides the resolution necessary to detect individual members of embedded clusters of early-type stars. A combination of Br$\gamma$ line observations with high-resolution radio maps provides extinction maps which allow the determination of spectral types of stars.

Our data demonstrate that both massive molecular clouds investigated by us contain clusters of OB stars, supporting the view that massive stars do not form in isolation. The object G11.11-0.40 is one of the rare cases where the central engine of an ultracompact H II region could be identified. This embedded O star is an excellent target for future near-infrared spectroscopy with very large telescopes in order to derive the exact spectral type of the object. There is some evidence that the ionizing source may even be a binary. The features of the object G341.21-0.21 suggest that it is in an early stage of massive star formation, presumably before the UCH II phase. Sensitive radio maps and molecular line data taken with interferometers are required to check this statement.

Acknowledgements

We would like to thank our referee, John Mathis, whose comments helped to improve the paper. B.St. acknowledges support from the DFG grant Ste 605/17-1.