... 1

Partially based on observations obtained at the European Southern Observatory, Paranal and La Silla, Chile (ESO 071.D-0151, 073.D-0327, 0.76.D-0037).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...

Table 1 and Appendix A are only available in electronic form at http://www.aanda.org

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... (TARA)

www.astro.psu.edu/xray/docs/TARA/

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... distribution

Note that the density of X-ray sources is low enough such that incompleteness effects caused by crowding do not influence the result, in contrast to studies of the spatial density of optical and/or near-IR sources.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...2005)

However, it should be noted that both estimates were determined from optical data alone, which would be entirely insensitive to a putative extended halo of low mass stars.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... (PIMMS)

http://heasarch.gsfc.nasa.gov/Tools/w3pimms.html

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...2006)

HST NICMOS data were also obtained for selected fields within the NTT/SofI fields, but the only counterparts to X-ray sources identified already had optical counterparts (Sect. 3.1) and hence these observations are not described further.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... stars

The effects of incompleteness due to crowding dominate the detection threshold for faint near-IR sources. As described in Brandner et al. (2007) this results in a significantly reduced sensitivity for the crowded inner regions of the cluster, which also host the majority of the X-ray point sources. Consequently, we conservatively choose to employ the 50% completeness limit of $K{\rm s}=16$ mag. found for the core region as the detection threshold for the NTT/SofI dataset.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... of 0.5''

For completeness, a cross correlation betwen the 2MASS and USNO-B1.0 datasets and the entire X-ray field was performed, utilising a search radius of 0.5'' within 5' of the aim point, and 1''beyond that offset. The results are presented in Table 1, but are not discussed further.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... associations

We note that randomly distributing either the X-ray or IR sources does not account for the obvious clustering in both, however including this clustering would increase the number of chance associations.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... WR K

The position of the X-ray source attributed by Skinner et al. (2006) is found to be $\sim$2.3'' to the West of the position of WR K in Crowther et al. (2006).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... WR U

Note that WRs #1 & #3 of Groh et al. (2006) are our WR U & W respectively.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...%

Of the 17 WR binary candidates, 11 are found to be photometrically variable by Bonanos (2007), whereas only one of the 7 apparently single WRs is found to be. We therefore speculate that the causes of the periodic and aperiodic variability in the WRs is related to binarity, possibly due to elipsoidal modulation and/or wind perturbation.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... limits

These were obtained by searching for additional X-ray sources at the optical co-ordinates of these stars utilising the aperture photometry technique described in Sect. 2.2. For the majority of stars this resulted in upper limits to the X-ray flux, although a few sources were detected at the $\sim$90% confidence level.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...2002)

The post shock temperature is given by T=(3/16)(m/k)v², where k is the Boltzmann constant, v the wind velocity, and m is the average particle mass. For a fully ionised plasma with solar abundances, and adopting m=10^-24 g, kT=0.05 keV for v=200 km s^-1 and 0.3 keV for 500 km s^-1.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... detected

Assuming log( $L_{\rm bol}$/ $L_{\odot}) \sim5.7$ for the YHGs, and $L_{\rm x}$/ $L_{\rm bol} \sim 10^{-6}$ yields $L_{\rm x} \sim 2\times10^{33}$ erg s^-1. For a 10⁷ K thermal plasma - as found for $\beta$ Dra by Ayres et al. (2005) - and D=5 kpc and $N_{\rm H}=2\times 10^{22}$ cm^-2, we may derive a detection limit of $\sim$ $5\times10^{31}$ erg s^-1 for our current observations.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... here

At an age of 4-5 Myr one would not expect Main Sequence stars earlier than $\sim$O7 V to be present in Wd1. Assuming $\log L/L_{\odot} \sim5.17$ for such stars (Crowther 2003) implies, via $L_{\rm x} \sim 10^{-7}~L_{\rm bol}$, a corresponding X-ray flux of $\sim$ $6\times10^{31}$ erg s^-1; consequently we do not expect to detect the OB Main Sequence population of Wd 1 in our current observations.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... core

Within the $5'\times 5'$ field previously considered we find 31 such sources.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... companion

The least massive star currently identified as a WR companion (Oskinova et al. 2005a).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... cycle

Recently, Negueruela et al. (2005) have identified a subset of transient supergiant HMXBs with extremely short duty cycles. Given observed upper limits to the quiescent X-ray flux of 10 ³²-10³³ erg s^-1, such a putative binary may have escaped detection if not undergoing a flare at the time of the observations. Consequently, the close similarity between the H$\alpha$ variability observed in W30a, and that seen for the SFXTs AX J1841.0-0536 and 1845.0-0433 (Negueruela et al. 2005) is particularly intriguing.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... observed

H  I, He  I, O  I, N  II, Fe  II, Ca  II, Mg  II, Ni  III, S  II, S  III and Ar  III.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... panel b)

We note that there is a significant scatter in magnitudes for stars of apparently identical spectral type. Differential reddening likely causes a $\sim$0.6 mag dispersion in the I band, but clearly cannot account for the full range of magnitudes. Other effects which may contribute are: (i) unresolved binarity (ii) uncertainties in the spectral classification and hence bolometic correction and (iii) a genuine scatter in the intrinsic magnitudes of stars of the same spectral type, possibly due to variations in rotational velocity (e.g. Meynet & Maeder 2003); a full account of these effects is beyond the scope of this paper.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.