The available observational information on the geometrical structure, emission line and continuous spectra, line profiles, radial velocities and linear polarization of Herbig-Haro objects is briefly reviewed. We emphasize the inhomogeneous structure of the “classical” Herbig-Haro objects and the appearance of small “condensations” with radii of ~300-900 a.u. The apparent paradox of the presence of “gaseous nebula type” as well as “reflection type” Herbig-Haro objects is discussed.
A purely empirical model of the regions of line formation is discussed. It shows that regions of low density (Ne ~103 cm-3) cover the space between the condensations and most of the volume of the condensations themselves. Only between 0.1 and 1% of the volume of the condensations is covered by a high density medium (Ne ~ 4 x 104 cm-3, N ~105 cm-3) which, however, contributes very strongly to the formation of the spectrum.
Different theoretical models for the line forming regions are discussed. We strongly favor the shock wave theory in which the emission lines are formed in the cooling regions of (running) shock waves. The general agreement between observations and the new calculations by Raymond is emphasized, and the few remaining discrepancies are discussed. The possibilities of explaining other properties of Herbig-Haro objects (including time scales, sizes and filling factors of condensations) are described.