The most reliable census of X-ray emitting T Tauri stars comes from a few deep ROSAT images of nearby active star forming clouds, in which the newly discovered X-ray emitting stars have been characterized spectroscopically and photometrically and placed on the Hertzsprung-Russell diagram. These include L1495W in Taurus-Auriga (Strom & Strom 1994), Chamaeleon I (Figure 9; Feigelson et al 1993, Lawson et al 1996) and IC348 in Perseus (Preibisch et al 1996, Herbig 1998). These X-ray surveys, combined with previous optical-infrared YSO surveys, show that Class III outnumber Class II stars within active star forming regions at least by factors of 2-3:1. Surprisingly, in the Chamaeleon I and Taurus-Auriga clouds, the age distribution of the Class III stars is substantially identical to the Class II age distribution (Figure 9).
The large population of X-ray discovered Class III stars has two immediate consequences. First, the star formation rate of a cloud, measured in stars/Myr, is higher than that inferred from early samples dominated by emission line Class II stars. The inferred star formation efficiency is probably more than doubled, but this quantity is more difficult to measure as older Class III stars may have drifted far from the cloud and be missing from the census. Also, molecular material is likely to decrease over time due to star formation and cloud disruption or evaporation. Nonetheless, it is likely that star forming clouds once thought to have star formation efficiencies around 10% (Cohen & Kuhi 1979) may actually have efficiencies of order 20-30%.
Second, the presence of many Class III stars near the stellar birthline implies that, at least in Chamaeleon I, about half of low-mass stars cease interacting with their circumstellar disks (and usually lose any detectable disk) within 1 Myr. Millimeter studies of CTT and WTT populations also indicate that early disk dissipation is common and can occur over timescales 6#6 yrs or less (André & Montmerle 1994). Earlier estimates that disks dissipate on times scales of 2-3 Myr based on less complete Einstein Observatory data (Strom et al 1989) are probably not reliable. Indeed, it is not clear that there is any preferred disk lifetime, as Class II and III stars coexist along most of the Hayashi track. Rather, the data suggest that the distribution of disk longevities ranges smoothly from 105 to 77#77 yrs (Lawson et al 1996). This result may be an important input into models of planet formation in the solar nebula and other circumstellar disks. This empirical evidence for a wide range of disk lifetimes agrees nicely with theoretical analyses indicating that star-disk coupling can efficiently transfer momentum from the star so that a broad distribution of disk lifetimes explains the wide disperson of rotational velocities among T Tauri stars and open cluster stars (Bouvier et al 1997).
An early unbiased survey of lithium-rich stars in a young stellar cluster suggests that X-ray selected samples are reasonably complete for intermediate-mass stars ten Myr or younger (Preibisch et al 1998b). If the census of all YSOs from a given cloud is complete, then the cloud's star formation history can be established by counting stars vertically along the Hayashi tracks. One such analysis in the Ophiuchus cloud complex suggests that star formation starts slowly and accelerates towards a terminal starburst (Martín et al 1999), as may be expected if ambipolar diffusion delays gravitational collapse (Palla & Galli 1997). However, the distribution of stellar ages in the Chamaeleon I cloud suggests a more continuous star formation history (Lawson et al 1996).