The bipolar outflow from L1551-IRS 5 is perhaps the archetypal example of-the outflow phenomenon among young stellar objects (YSOs). In this paper we report on near-infrared observations of the L1551-IRS 5 region in the 2.12 mu m line of-shocked H-2. Although copious amounts of emission are discovered in the blueshifted molecular lobe, surprisingly, none is found in the redshifted one. Generally, the H-2 emission is knotty or amorphous. Only one H-2 emission region, HH 260, appears bow shock shaped in our data, with the axis of the bow pointing back towards IRS 5. Our H-2 observations are compared with high quality optical data obtained by us through [S II] lambda lambda 6716, 6731, H alpha, and nearby narrowband continuum filters. There appears to be only a weak spatial correlation between the optical emission, recording shocks with V-shock greater than or equal to 40 km s(-1), and the shocked H-2 emission, which traces lower shock velocities. As an example, the IRS 5 jet is not detected at 2.12 mu m. Much of the optical and near-infrared line emission is patchy, and covers a wide range of position angles as seen from IRS 5. All of this emission, however, is confined to an ovoid-shaped and presumably wind-blown cavity which has a total extent of 0.45 pc and a dynamical age of about 2200 yr. Within this emission-line cavity, which we refer to as the ''H alpha cavity,'' the high-velocity CO gas is also found. The slower CO gas and an additional cavity seen only in scattered light lie outside this emission Line cavity. The presence of both the highly collimated jet from IRS 5 and the poorly collimated flow seen much further away [as evidenced by the Herbig-Haro (HH) emission within the Ha cavity spread over a large range of position angles as seen from the source] is difficult to reconcile. In the past it has been suggested that there is a poorly collimated wind from the source in addition to the jet and that it is this component, rather than the jet, that drives the associated CO outflow and powers the large amount of ''off-axis'' HH emission. This idea, however, does not account for many of the observations in IRS 5 and in other poorly collimated outflows. We therefore suggest that the opening angle of the outflow varies with time so that sources like IRS 5 pass through alternative phases of poor and high degrees of collimation. In IRS 5, this scenario can in principle explain the observed ''scattered light cavity'' and shell of low-velocity CO gas, both of which enclose the ''H alpha cavity'' and the high-velocity CO flow. The implications for other poorly collimated outflows from YSOs, and for the acceleration of their CO flows, is also examined. Finally, we briefly consider possible implications for the origin of cometary reflection nebulae. (C) 1995 American Astronomical Society.
