Abstract

We have surveyed 84 Class 0, Class I, and flat-spectrum protostars in mid-infrared [Si II], [Fe II], and [S I] line emission, and 11 of these in far-infrared [O I] emission. We use the results to derive their mass. outflow rates, (M) over dot(w). Thereby we observe a strong correlation of (M) over dot(w) with bolometric luminosity, and with the inferred mass accretion rates of the central objects, (M) over dot(a), which continues through the Class 0 range the trend observed in Class II young stellar objects. Along this trend from large to small mass. flow rates, the different classes of young stellar objects lie in the sequence Class 0-Class I/flat-spectrum-Class II, indicating that the trend is an evolutionary sequence in which (M) over dot(a) and (M) over dot(w) decrease together with increasing age, while maintaining rough proportionality. The survey results include two that. are key tests of magnetocentrifugal outflow-acceleration mechanisms: the distribution of the outflow/accretion branching ratio b = (M) over dot(w)/(M) over dot(a), and limits on the distribution of outflow speeds. Neither rules out any of the three leading outflow-acceleration, angular-momentum-ejection mechanisms, but they provide some evidence that disk winds and accretion-powered stellar winds (APSWs) operate in many protostars. An upper edge observed in the branching-ratio distribution is consistent with the upper bound of b = 0.6 found in models of APSWs, and a large fraction (31%) of the sample have a. branching ratio sufficiently small that only disk winds, launched on scales as large as several au, have been demonstrated to account for them.