Assuming normal interstellar conditions (Bohlin et al. 1978) and using average extinctions for typical 1600-6400 au wide filaments we derive molecular hydrogen densities of the order of . This is not inconsistent with upper limits of about obtained by Schnepps et al. (1980), since their beam width is comparable to the trunk diameter. They also derived excitation temperatures for CO of 5-20 K for the trunk, which would imply internal gas pressures up to a few times . Using the derived gas densities and diameters of the filaments we find that the filaments have linear mass densities that are several orders of magnitude below what is needed for self-gravitation to balance the internal pressure. If we assume that the filaments are close to hydrostatic equilibrium, what forces are then responsible for maintaining the structures?
If the elephant trunk is engulfed in the expanding H II region with and an electron density of at from the central cluster (Menon 1962), the outer kinetic pressure is while the dynamic pressure is . The background radiation field is dominated by photons from the central cluster with a total luminosity of (Cox et al. 1990), the corresponding radiation pressure on the front of the trunk being . Hence, adding up the three components yields . This is sufficient to balance even the denser filaments discussed above, without including effects from pervasive stellar winds, and sufficient to exert drag forces on the trunk as a whole. From the velocity gradient along the trunk of , as observed in CO, Schneps et al. (1980) also concluded that there is a history of stretching of the trunk under the influence of the H II region.
The external pressure is thus sufficient to confine the filaments. This, however, does not explain their helical structure. In fact, it seems very difficult to devise any physical process in the confining medium that might explain the helical complexity of the trunk. We find a primordial alignment with a helical magnetic field to be the most compelling explanation for the observed structure. It is to be noted that a helically shaped magnetic field presents direct evidence of the existence of electric currents. Moreover, the observed helical morphology strongly suggests that the intrinsic magnetic field also acts as the unifying agent responsible for maintaining the trunk-system. This would explain the velocity gradient observed in the trunk. The northern part of the trunk connects to a part of the structure running perpendicular to the direction to NGC 2244 which absorbs a larger part of the momentum of stellar radiation and winds from the central cluster than does the southern tip of the trunk.
A close inspection of the trunk image reveals that the individual filaments seem to be well separated. In order for the electromagnetic forces not to be too strong and disruptive the electric current paths in the trunk are likely to form a nearly force-free geometry, as is e.g. often found in some solar prominences. By numerical computations we have studied current configurations consisting of 4-7 helical cables situated on the surface of a cylinder; an adequate model of the observed filament system. All the cables are supposed to carry currents that are of the same magnitude. These cable configurations are found to be force-free, i.e. the magnetic field generated by the currents is aligned with the cables, when the wavelength of the helices is about 7-9 times the cylinder radius (Carlqvist 1998). This agrees very well with the observations. Ensembles of helical filaments may be described as magnetic flux ropes and Cox (1996) envisioned that such flux ropes could control extremely dense small-scale H I features.
Flows of ionized gas around primordial density inhomogeneities (Pottasch 1956) or dynamical instabilities occurring when ionized gas penetrates a clumpy neutral medium (Schneps et al. 1980; Spitzer 1954; Garcia-Segura & Franco 1996) have been discussed in connection with the formation of elephant trunks. However, a worm-like network of globules found outside the H II region led Block et al. (1992) to suggest that diffuse UV radiation from the OB stars plays a role in figuring primordial clumps. Furthermore, radiation-controlled "etching" and implosion of cloudlets have been considered (Patel et al. 1993; White et al. 1997). Although these scenarios might provide possible explanations for trunks observed in other H II regions, neither manages to explain the morphology of the trunk system observed in the Rosette nebula. They are in conflict with the observed well-confined, rope-like appearance of the interconnected trunks, which partly are oriented tangentially around a central cluster (Fig. 1b).
© European Southern Observatory (ESO) 1998
Online publication: March 10, 1998