There is now growing evidence that the cosmic jet phenomenon manifests itself in a remarkable way in regions of active star formation embedded in dense molecular clouds. The first indications for oppositely directed, supersonic outflows from young stars were provided by molecular line observations (most notably of CO) which detected spatially separated regions of redshifted and blueshifted emission in association with embedded infrared sources. About twenty sources of this kind have been identified so far, and more are continuously being discovered; they typically have radii ∼1018 cm, velocities ∼10–50 km s−1, dynamical ages ∼104 yr, and energies ∼1046-1047 erg s−1 (see Bally and Lada 1983 for a review). Statistical arguments indicate that energetic outflows of this type are probably a common feature in stellar evolution, and that they occur in both massive and low-mass stars. Direct evidence that the outflows in many cases are highly collimated was subsequently provided by the detection of high-velocity Herbig-Haro objects (optical emission clumps with typical masses ∼10−5M⊙) along the axes of the bipolar CO lobes. Proper-motion measurements are now available for a number of these objects (e.g., Herbig and Jones 1981), and they invariably reveal that the velocity vectors (of typical magnitudes 200–400 km s−1) point away from the central star. The clumps are often found to consist of many sub-condensations which move independently with disparate speeds, but which nevertheless travel in the same general direction with an angular spread ≲ 10°. Finally, radio continuum observations (e.g., Cohen et al. 1982) and deep CCD images (e.g., Mundt and Fried 1983) have shown that the collimation of the outflows is already well established on scales of ≲ 1015 cm.