2 The ONC: An update

Figure 1: Contour plot of the logarithm of the conversion factor (CF) from ROSAT HRI count-rates to unabsorbed flux, in the 0.1-4.0 keV spectral band, for coronal thermal sources (Raymond-Smith emission model) as a function of source temperature, kT, and absorption, $N_{\rm H}$. Lines refer to loci in which Log(CF) is constant and equal to the value reported in the corresponding label plus -10.07, i.e. the Log(CF) corresponding to typical parameter values: kT=2.16, $N_{\rm H}=2\times 10^{21}$ (marked in the figure by a triangle). Line spacing is 0.1 dex in Log(CF). The ordinate axis on the right gives the standard conversion between $N_{\rm H}$ and optical absorption, AV: $N_{\rm H}=2\times 10^{21}A_{V}$.

Our analysis method throughout this paper is based on the work of Flaccomio et al. (2002a,b) on the ONC. This cluster is arguably the best target available for our study because we have access to a rich and well characterized sample of members spanning a wide range of masses. We refer the reader to Flaccomio et al. (2002a,b) for a full description of the X-ray and optical data used here. The extinction limited sample ( A_V < 3.0) discussed in this latter work has little field contamination and is complete almost down to the lowest stellar masses. Using the Ca II line ( $\lambda = 8542$ Å) as an indicator for circumstellar accretion, Flaccomio et al. (2002b) obtained with high statistical significance the result that low mass stars ( $M \lesssim 3~ M_\odot$) with this line in strong emission (EW < -1) have systematically lower $L_{\rm X}$ and $L_{\rm X}/L_{\rm bol}$ values respect to stars with the line in absorption (EW > 1).

Here we state that an analogous result is obtained, albeit with smaller significance ( $2.5{-}3 \sigma$), comparing the $L_{\rm X}$ and $L_{\rm X}/L_{\rm bol}$distributions of stars with large and small near IR excess ( $\Delta(I-K) > 0.8$and $\Delta(I-K) < 0.4$ respectively). The X-ray and optical/IR data are presented in Flaccomio et al. (2002a) and the $\Delta(I-K)$ values are taken from Hillenbrand et al. (1998). Figures 2 and 3 shows the maximum likelihood $L_{\rm X}$ and $L_{\rm X}/L_{\rm bol}$ distributions for these two classes of stars in 6 different mass bins. The range of confidence with which we can exclude that the two distributions are randomly extracted from the same parent population, according to the tests in the ASURV package (Feigelson & Nelson 1985), is given inside each panel.

Figure 2: X-ray luminosity functions for stars with high- and low-NIR excess (solid and dashed lines, respectively) in the Orion Nebula Cluster. Panels refer to different mass ranges, as indicated by legends. Also reported are the numbers of detected (d) and undetected stars (u) used for XLFs of high- and low-accretion subsamples and the $\sigma$-equivalent significance range for the difference between the two distributions (see text).

Figure 3: Same as Fig. 2 for $L_{\rm X}/L_{\rm bol}$.

We also note that very similar results are obtained, both with the Ca II line and with $\Delta(I-K)$ as a discriminant, if only X-ray detected members are considered in the distribution functions. As reminded in the introduction, these latter results exclude that the difference in the distributions is due to the preferential selection of faint CTTS (which is anyway not expected given that the sample is not selected from either accretion or disk indicators).