The All-Sky Survey performed by the German X-ray satellite ROSAT offers for the first time the opportunity to study a spatially complete sample of X-ray sources in any star forming region. For a detailed description of ROSAT and its detectors we refer to Trümper (1983) and Pfeffermann et al. (1986). During the RASS the sky was scanned in great circles, using the Positional Sensitive Proportional Counter (PSPC). Each object in the sky was observed during several scans once every 93 minutes for up to about 30 s per scan. The number of scans per object depends on the ecliptic latitude. In the Lupus area each point in the sky was observed during some 30 scans, giving a mean exposure time of 406 s with an rms scatter of 42 s.
The RASS data were reduced using the EXSAS
(EXtended Scientific Analysis System) software
package developed at MPE, Garching.
Source detection
was performed seperately in
five different energy bands - broad (
), soft (
), hard (
), hard1 (
) and
hard2 (
). After application of both local and map source
detection algorithms detected sources were merged and tested with a maximum
likelihood technique (Cruddace et al. 1988).
We used a treshold of 
for
the likelihood of existence ML, which is defined as 
(where P is the probability for the existence of a source). This procedure
resulted in the detection of 437 sources.
For the selection of candidate sources we first searched the SIMBAD database for optical counterparts of the X-ray sources. We excluded all sources that could be identified with known objects within 60'' from the X-ray position. (The positional accuracy of the RASS is discussed, e.g., by Motch et al. 1991 and Neuhäuser et al. 1995b. We decided to use a rather conservative, i.e. large error circle.)
Catalogued stars, for which a TTS nature
could neither be ascertained nor rejected on the basis of available data,
were kept in the list for
further study.
Next we visually inspected
the ESO SRC(R) plates to exclude
all sources without a stellar object of 
(which
converts to 
for an M2V star) inside
the aforementioned error radius.
This resulted in a final list of 298 candidate sources for
spectroscopic follow-up observations.
Unlike in our study of the Taurus SFR (Wichmann et al. 1996), no
selection of candidate objects by means of the X-ray energy distribution
has been performed.
Table 1: Pointed observations in Lupus 1
Table 2: Pointed observations in Lupus 2
Table 3: Pointed observations in Lupus 3