The detector used for the XRT survey was the Position Sensitive Proportional Counter
(PSPC) which offers a spatial resolution of 
20
and medium energy
resolution (40% at 1keV) in the range of 0.07 to 2.4keV (Pfeffermann et
al. 1986). During survey observations, the sky was scanned in great
circles bound to be perpendicular to the direction to the Sun thus covering the
whole celestial sphere in six months. A given part of the sky was seen several
times at 96min intervals corresponding to the ROSAT orbit and scan period. For a
given position on the sky, the number of available scans and total exposure time
depends on the ecliptic latitude, the poles being the most deeply observed, and on
possible detector switch off episodes (caused by radiation belts) close to the
earth magnetic poles and South Atlantic Anomaly. With the field of view of 2
of the PSPC, each source was seen during a maximum of 32s each orbit. Because
of the blur at large off axis angles, the average survey point spread function
(FWHM 
3
) is significantly larger than the one on-axis but still
allows source positioning with a 1 
accuracy better than 20
\
(Voges 1992).
The area chosen for our investigations was a rectangle extending from l = 86
to
94
and from b = -5
to +5
. However, part of the
selected X-ray region is contaminated by hard diffuse emission from the Cygnus
super bubble (Cash et al. 1980) and after rejection of the high
background areas the final size of the region was 64.5deg
. Actual area
boundaries are described in detail in Motch et al. (1997). By
chance, the selected region is located in a range of galactic longitude which is
closest to the north ecliptic pole. A strong gradient in exposure time is present
in the selected field from about 500s at 
to 1000s at 
. However, 84% of the sources in our field have exposure times ranging from
700 to 900s.
The final list of sources was extracted from the merged survey data using the dedicated
Extended Scientific Analysis System (EXSAS) developed at MPE (Zimmermann et
al. 1992). Source detection was run separately in three energy bands
(soft, 
; hard, 
and broad, 
)
using the maximum likelihood (ML) algorithm and accepting sources above ML = 7.
The three source lists were then merged into a single one in which we defined the
``merged" source count rate and ``merged" maximum likelihood as being the maximum of
the count rates and maximum likelihood respectively in all three bands. The
retained X-ray coordinates were those of the broad band detection or those of the
hard or soft bands when the source was not detected in the broad band. A total of
95 and 128 sources were detected above ML = 10 and 8 respectively and 30
additional sources had ML in the range of 7 to 8. Each of the sources with 
has an associated index running from 1 to 128, increasing with decreasing
count rates. Because of the expected high fraction of spurious sources with 7
we do not discuss these marginal detections in detail. We list
separately these doubtful detections which were given index numbers in the range
of 129 to 158.
The 90% confidence error radius was estimated assuming that the uncertainty on the source
localization could be expressed as the quadratic sum of the statistical error determined
by the maximum likelihood detection algorithm and of a systematic bore sight error of
8
. For a subset of sources, standard ROSAT hardness ratios were computed.
We defined hardness ratios 1 and 2 with energy boundaries as used in the latest SASS
versions:
where (A-B) is the raw background corrected source count rate in the
A-B energy range expressed in keV. Tables 2 (click here) and 3 (click here) summarize the
main X-ray properties of each detected ROSAT source. From the bending of the
\
curve we estimate that at ML = 10 our completeness level is 0.02cts s
which
roughly corresponds to a flux limit of 2 10
erg cm
s
. At this flux level
the source density is 
1.0 source deg
. Using ML = 8 may allow
to improve on the sensitivity (see Motch et al. 1997) and at an
estimated completeness level of 0.012cts s
the source density is 
1.8
deg
. However, the enhanced number of spurious sources complicates the
definition of a completeness level at ML = 8.