The various and hitherto partially unsolved problems relative to the o
rigin of bipolar jets or highly collimated fast outflows in planetary
nebulae are reviewed within the framework of a stationary magnetohydro
dynamic model. In order to explain the observations of high polar velo
cities and the presence of polar blobs or knots in planetary nebulae,
theoretical models are proposed taking into account both a large scale
azimuthal magnetic field and an anisotropic turbulent velocity field.
The models predict equatorial-to-polar density ratios which are rathe
r small, in the range 2 to 3. Conversely, the polar-to-equatorial velo
city contrasts are higher, with typical values upto 10. Thus the ad ho
c hypothesis implicit in the literature that the density contrast is v
arying in inverse ratio to the velocity one, does not seem well adapte
d to the bipolar jet phenomenon in planetary nebulae. We point out, th
erefore, that the bipolar jets have to be considered as a transient as
pect of a very complex phenomenon. The model can be applied to objects
such as He 2-104 or Mz3, M2-9.