A&A 439, 559-563 (2005)
DOI: 10.1051/0004-6361:20052886

Eclipsing binaries in the Galactic bulge: candidates for distance estimates^,

M. A. T. Groenewegen

Instituut voor Sterrenkunde, K.U. Leuven, PACS-ICC, Celestijnenlaan 200B, 3001 Leuven, Belgium

Received 16 February 2005 / Accepted 26 April 2005

Abstract
The 222 000 I-band light curves of variable stars detected by the OGLE- II survey in the direction of the Galactic Bulge have been searched for eclipsing binaries (EBs). A previously developed code to analyze lightcurve shapes and identify long period variables (LPVs) has been adapted to identify EBs. The parameters in the modified code have been optimised to recover a list of about 140 detached EBs in the Small Magellanic Cloud previously identified in the literature as particularly well suited for distance estimates (and wich have periods $\ga$ 0.85 days). The power of the code is demonstrated by identifying 16 and 178 previously uncatalogued EBs in the SMC and LMC, respectively. Among the 222 000 variable stars in the direction of the Galactic Bulge 3053 EBs have been identified. Periods and phased lightcurves are presented.

Key words: stars: distances - binaries: eclipsing - Galaxy: center

1 Introduction

Detached double-lined eclipsing binaries (hereafter EBs for short) are very suitable as primary distance indicators when accurate photometry and radial velocity (RV) data are combined (e.g. Andersen 1991. The massive photometric monitoring data obtained in the course of the micro lensing surveys in the 1990's of the Small and Large Magellanic Clouds has revealed many candidate detached eclipsing binaries based on the shape of the lightcurves (Grison et al. 1995; Alcock et al. 1997; Udalski et al. 1998; Bayne et al. 2002; Wyrzykowski et al. 2003, 2004).

Some of them have been followed-up with spectroscopy to obtain RV data and perform a combined lightcurve and RV curve analysis to obtain the distance to the system (see e.g., Guinan et al. 1998; Hilditch et al. 2005; and the review by Clausen 2004, for a list of systems with recent distance estimates). What makes EBs particularly powerful is that the error in the distance estimate in a single well-observed system (0.06-0.10 in distance modulus, see Clausen 2004) is already comparable to the dispersion in optical and infrared Cepheid PL-relation based on hundreds of stars (e.g. Sandage et al. 2004; Nikolaev et al. 2003).

In the present paper a list of eclipsing binaries in the direction of the Galactic Bulge (GB) is presented based on analysis of OGLE- II data. Previously, 1650 EBs have been identified in OGLE- I data (Udalski et al. 1994, 1995a,b, 1996, 1997).

The paper is structured as follows. In Sect. 2 the OGLE- II surveys is briefly described. In Sect. 3 the model to identify EBs from the lightcurve shape is briefly presented. The results are presented in Sect. 4, and discussed in Sect. 5.

2 The data sets

The OGLE- II micro lensing experiment observed fourty-nine fields in the direction of the GB. Each field has a size 14.2 $^\prime$ $\times$ 57 $^\prime$ and was observed in $\it BVI$ , with an absolute photometric accuracy of 0.01-0.02 mag (Udalski et al. 2002). Table 1 lists the galactic coordinates of the field centers and the total number of sources detected in these fields.

Wozniak et al. (2002) present a catalog of about 222 000 variable objects based on the OGLE observations covering 1997-1999, applying the difference image analysis (DIA) technique on the I-band data. The data files containing the I-band data of the candidate variable stars was downloaded from the OGLE homepage (http://sirius.astrouw.edu.pl/~ogle/).

Table 1: Properties of the OGLE-fields and the number of detected eclipsing binaries.

3 Analysis of the lightcurve shape and selecting eclipsing binaries

The model to analyse the lightcurve shape and identify long period variables (LPVs) is described in detail in Appendices A-C in Groenewegen (2004, herafter G04).

Briefly, a first code (see for details Appendix A in G04) sequentially reads in the I-band data for the objects, determines periods through Fourier analysis, fits sine and cosine functions to the light curve through linear least-squares fitting and makes the final correlation with the pre-prepared DENIS and 2MASS source lists. All the relevant output quantities are written to file. This part of the code is adapted significantly to better deal with the specific properties of eclipsing binaries w.r.t. pulsational variables, as described in Appendix A of the present paper. Brief, phase dispersion minimization is used next to Fourier analysis to select the true orbital period, and the EBs candidates are selected based on statistical properties of the phased light curve.

The output file of the first code is read in by the second code (see for details Appendix B in G04). A further selection may be applied (typically on period, amplitude and mean I-magnitude), multiple entries are filtered out (i.e. objects that appear in different OGLE fields), and a correlation is made with pre-prepared lists of objects for cross-identification. The output of the second code is a list with EB candidates.

The third step (for details see Appendix C in G04) consists of a visual inspection of the fits to the (phased) light curves of the candidate EBs and a literature study through a correlation with the SIMBAD database. Non-EBs are removed, and sometimes the fitting is redone. The final list of EB candidates is compiled.

4 Results

4.1 Testing the code on selected SMC data

In order to set and finetune the parameters described in Appendix A that define the selection of an EB, and also to demonstrate that with these particular settings EBs can indeed be retrieved, the code is tested and run on the OGLE data of variable stars in the Magellanic Clouds (hereafter MCs; Zebrun et al. 2001).

As the ultimate aim of the project is to define EBs in the GB suitable for distance determinations the parameters that define the selection of EBs (3 statistical parameters that define the shape of the phased lightcurve, and the parameters significance and hifac of the NUMERICAL RECIPES (Press et al. 1992) subroutine fasper) are tuned to retrieve the stars in the lists of "ideal'' distance indicators in the SMC previously selected by Udalski et al. (1998) and Wyithe & Wilson (2001) based on the first release of OGLE data (based on a shorter timespan of OGLE data and aperture photometry instead of DIA). What constitutes an "ideal'' EB in terms of a distance indicator is difficult to quantify exactly. Typically the system should be well-detached with small values of the stellar radii relative to the semi-major axis ( $r/a \la 0.2)$ and have deep eclipses (e.g. Wyithe & Wilson 2001).

Not all of the stars from Udalski et al. (1998) and Wyithe & Wilson (2001) are listed in the Zebrun et al. (2001) data set. In fact, Wyrzykowski et al. (2004) state that of the 1527 EBs listed by Udalski et al. (1998) only 935 can be found in the DIA catalog of variable stars by Zebrun et al. (2001), and quote the incompleteness of the DIA catalog for faint stars as the probable reason. Of the 153 stars listed by Udalksi et al. (1998) and the 22 unique additional objects in Tables 4 and 5 of Wyithe & Wilson (i.e. not already listed in Udalski et al.), 142 are found back by Zebrun et al. (2001), and these stars represent the test data set.

Of these 142, 137 are correctly identified as EBs by the program using the parameters mentioned in Appendix A. One object (OGLE005102.88-730941.0) is found with the period as listed by Udalski et al. (1998) but the phased light curve is poor with the DIA data set, and the object is not suitable as a distance indicator from the present data. Two stars would have been correctly identified as EBs were it not that the significance of the peak of the Fourier spectrum is above the adopted threshold. The other two are missed because the parameters that are being used to describe the shape of the phased lightcurve are outliers. At the end of the day, it is deemed acceptable to miss 4 stars out of a 141 suitable candidates. It should be mentioned that these "ideal'' distance indicators have orbital periods larger than 0.85 days and so the code is optimised to be sensitive to this period range, and hence will be biased against shorter orbital periods.

4.2 Applying the code to SMC and LMC data

As a further test, the code is applied to the full set of 68 000 lightcurves of variable stars in the SMC and LMC by Zebrun et al. (2001), and compared to the outcome of the previous work by Wyrzykowski et al. (2003, 2004), based on the same dataset as is the present work.

Wyrzykowski et al. (2003) find 2580 EBs in the LMC, and Wyrzykowski et al. (2004) find 1351 EBs in the SMC, based on an artificial neural network. Note that they attempted to find all types of eclipsing binaries, while the aim of the present paper is to detect EBs potentially suitable for distance determination, hence preferentially detached systems.

Applying the code resulted in 1856 LMC and 752 SMC candidate EBs.

For the SMC, 714 of the 752 objects are listed in Wyrzykowski et al. (2004) as EBs. Of the 38 not listed in Wyrzykowski et al., 20 are eliminated at the visual inspection stage as the phased lightcurves are not typical of that of an EB; 13 are in fact known Cepheids. Of the 18 remaining stars, 2 are classified as EB by MOA (Bayne et al. 2002), and 16 are new EBs not previously identified. Names, periods and the phased lightcurve are presented in Appendix B.

For the LMC, 1616 of the 1856 objects are listed in Wyrzykowski et al. (2003) as EBs. Of the 240 not listed in Wyrzykowski et al., 51 are eliminated at the visual inspection stage as the phased lightcurves are not typical of that of an EB; 7 are known Cepheids, 2 are known RR Lyrae and 24 are LPVs. Of the 189 remaining stars, 11 are classified as EB by MACHO (Alcock et al. 1997), and 178 are new EBs not previously identified. Names, periods and the phased lightcurve are presented in Appendix B.

Table 2: First entries in the electronically available table, which lists: OGLE-name, oribital period, and time used to define phase zero.

$\begin{figure} \par\includegraphics[width=17.4cm,clip]{paper_bulge.ps}\end{figure}$	Figure 1: First 15 entries of the electronically available figure with all phased lightcurves of eclipsing binaries in the direction of the Galactic Bulge. Crosses indicate data points not included in the analysis. Phase zero is given by the Julian Date listed in Table 2.
Open with DEXTER

The conclusion is that the simple method developed for detecting EBs is complementary to the neural network method used in Wyrzykowski et al. (2003, 2004) as, considering both MCs, about 7.5% of the detected objects are newly discovered EBs. The number of false candidates (mostly cepheids, RR Lyrae and LPVs) is about 2.5%.

4.3 Applying the code to the Galactic Bulge

The numerical code is applied to the OGLE- II data in the direction of the Galactic Bulge. After visual inspection of the lightcurves a sample of 3053 objects remain. The largest number of objects that are removed at this stage are LPVs especially at longer periods ( $P_{\rm binary}$ $\ga$ 150, respectively $P_{\rm pulsation}$ $\ga$ 75 days) and EBs with orbital periods $\la$ 0.85 days where an alias frequency is picked up, which nevertheless phases well enough to fulfill the initial selection criteria.

The number of objects per field is listed in the last column of Table 1. Table 2 lists the 3053 objects with periods, and Fig. 1 displays the phased lightcurves. The typical error in the period is 3 $\times$ 10^-4 P. Only one object is listed in the SIMBAD database, namely bul_sc26_3510 (RA = $17{\rm h}47{\rm m}28.25{\rm s}$ , Dec = $-34{\rm d}46{\rm m}31.0{\rm s}$ ), a.k.a. MM5-A V47 in Udalski et al. (1997) and classified as a W UMa-type binary. The period quoted there is 1.10577 d while in the present analysis 1.105801 d is derived.

5 Summary

The paper presents a simple method to detect eclipsing binaries based on the shape of the phased lightcurve. In view of the ultimate aim to use EBs to derive the distance to the Galactic Centre, the parameters of the model are finetuned to retrieve a large sample of EBs in the SMC previously classified in the literature as potentially suitable for distance determinations. The code is run on the 68 000 OGLE variable stars in LMC and SMC to find 194 apparently previously unclassified EBs, and on the about 222 000 OGLE variable stars in the direction of the Bulge to find about 3000 EBs. It should be stressed that due to the finetuning of the model no short period EBs ( $P \la 0.85~{\rm d}$ ) are among these.

It should also be stressed that altough the code was tuned to retrieve EBs peviously classified in the literature as suitable for distance determinations, the converse is not necesarily true: the objects returned by the code and listed in Tables 2 and B.1 are not automatically suitable as potential distance indicators. For example, none of the 15 objects displayed in Fig. B.1 seems suitable for this purpose. In this case this is not surprising as only newly discovered EBs, not in the exhaustive lists of Wyrzykowski et al. (2003, 2004), are listed and so these objects have relatively low S/N lightcurves. Of the lightcurves displayed in Fig. 1 only a few appear suitable for follow-up studies.

In the future it is planned to select a small sub-sample among the 3000 objects based on a proper analysis of the lightcurve in terms of radii determination and eclipse depth and obtain multi-epoch spectroscopic data in order to derive the fundamental parameters (mass, radius, effective temperature, metallicity) and distance.

Acknowledgements

This research has made use of the SIMBAD database, operated at CDS, Strasbourg, France.

References

Alcock, C., Allsman, R. A., Alves, D., et al. 1997, AJ, 114, 326 [NASA ADS] [CrossRef] (In the text)
Andersen, J. 1991, A&AR, 3, 91 [NASA ADS]
Bayne, G., Tobin, W., Pritchard, J. D., et al. 2002, MNRAS, 331, 609 [NASA ADS] [CrossRef] (In the text)
Clausen, J. V. 2004, NewAR, 48, 679 [NASA ADS] (In the text)
Grison, P., Beaulieu, J.-P., Pritchard, J. D., et al. 1995, A&AS, 109, 447 [NASA ADS] (In the text)
Groenewegen, M. A. T. 2004, A&A, 425, 595 [EDP Sciences] [NASA ADS] [CrossRef] (G04) (In the text)
Groenewegen, M. A. T., & Blommaert, J. A. D. L. 2005, A&A, submitted (In the text)
Guinan, E. F., Fitzpatrick, E. L., DeWarf, L. E., et al. 1998, ApJ, 509, L21 [NASA ADS] [CrossRef] (In the text)
Hilditch, R. W., Howarth, I. D., & Harries, T. J. 2005, MNRAS, 357, 304 [NASA ADS] (In the text)
Nikolaev, S., Drake, A. J., Keller, S. C., et al. 2003, ApJ, 601, 260 [NASA ADS] (In the text)
Popowski, P., Cook, K. H., & Becker, A. C. 2003, AJ, 126, 2910 [NASA ADS] [CrossRef]
Press, W. H., Teukolsky, S. A., Vetterling, W. T., & Flannery, B. P. 1992, Numerical Recipes in Fortran 77 (Cambridge U. P.) (In the text)
Sandage, A., Tammann, G. A., & Reindl, B. 2004, A&A, 424, 43 [EDP Sciences] [NASA ADS] [CrossRef] (In the text)
Schultheis, M., Ganesh, S., Simin, G., et al. 1999, A&A, 349, L69 [NASA ADS]
Stellingwerf, R. F. 1978, ApJ, 224, 953 [NASA ADS] [CrossRef] (In the text)
Sumi, T. 2004, MNRAS, 349, 193 [NASA ADS] [CrossRef]
Udalski, A., Kubiak, M., Szymanski, M., et al. 1994, AcA, 44, 317 [NASA ADS] (In the text)
Udalski, A., Szymanski, M., Kaluzny, J., et al. 1995a, AcA, 45, 1 [NASA ADS] (In the text)
Udalski, A., Olech, A., Szymanski, M., et al. 1995b, AcA, 45, 433 [NASA ADS]
Udalski, A., Olech, A., Szymanski, M., et al. 1996, AcA, 46, 51 [NASA ADS] (In the text)
Udalski, A., Olech, A., Szymanski, M., et al. 1997, AcA, 47, 1 [NASA ADS] (In the text)
Udalski, A., Soszynski, I., Szymanski, M., et al. 1998, AcA, 48, 563 [NASA ADS] (In the text)
Udalski, A., Szymanski, M., Kubiak, M., et al. 2002, AcA, 52, 217 [NASA ADS] (In the text)
Wozniak, P. R., Udalski, A., Szymanski, M., et al. 2002, AcA, 52, 129 [NASA ADS] (In the text)
Wyrzykowski, L., Udalski, A., Kubiak, M., et al. 2003, AcA, 53, 1 [NASA ADS] (In the text)
Wyrzykowski, L., Udalski, A., Kubiak, M., et al. 2004, AcA, 54, 1 [NASA ADS] (In the text)
Wyithe, J. S. B., & Wilson, R. E. 2001, ApJ, 559, 260 [NASA ADS] [CrossRef] (In the text)
Zebrun, K., Soszynski, I., Wozniak, P., et al. 2001, AcA, 51, 317 [NASA ADS] (In the text)

Online Material

Appendix A: The model

Compared to the code described in G04 the part of fitting sine and cosine functions to the lightcurves is omitted as the shape of the lightcurve of an EB is not described by such a functional form.

$\begin{figure} \par\includegraphics[width=8.8cm,clip]{EB_example_of_finding.ps}\... ...mm} \includegraphics[width=8.8cm,clip]{EB_example_of_not_finding.ps}\end{figure}$

Figure A.1: Example of a phased lightcurve of an eclipsing binaries ( left) and an long period variables ( right), and the two magnitude limits discussed in the text. The identification of an eclipsing binary is based on the number and statistical properties of the data points fainter than these two limits.

The determination of the period is done in two steps: using Fourier analysis (as in G04), and subsequently phase dispersion minimization (PDM), as described by Stellingwerf (1978), and which is new to the code.

The Fourier analysis is done with the NUMERICAL RECIPES (Press et al. 1992) subroutine fasper. Inputs to it are the time and magnitude arrays. In addition one has to specify two parameters, ofac and hifac, that indicate a "typical oversampling factor'' and the maximum frequency in terms of a "typical'' Nyquist frequency.

In the present work ofac = 22 (as in G04) and hifac = 21.0 are used (compared to 0.8 in G04). The latter parameter is the determining factor in both the computational speed, and the shortest period that can be found. In this case, it is set to correctly identify the main period in OGLE sc8/119647 (OGLE005938.46-730002.2) with a known orbital period of 0.85494 days (Wyithe & Wilson 2001), which is the shortest period in the test dataset described in the main text. It implies that in the configuration used in the present paper there is a bias in the detection of orbital periods shorter than about 0.85 days, of no severe consequence as the focus is finding detached EBs.

The outputs of fasper are the frequency where the peak of the power spectrum occurs and a number indicating a significance. One of the main parameters in the code is to provide the critical cut-off above which a period is not considered to be significant. In the present work significance = 2.2 $\times$ 10^-5 is used, and this is determined empirically, by using the set of detached EBs which the code should be able to retrieve (see below).

As had already been noted in G04 and Groenewegen & Blommaert (2005) in analysing LPVs, the frequency returned from the Fourier analysis in some cases turns out to be a harmonic rather than the true period and this needs then manual refitting of the lightcurve. The same turns out to be the case when the test data set of EBs was initially being analysed. Therefore it was decided to perform a PDM analysis at selected frequencies to decrease manual intervention and hence to allow for a more fully automatic analysis. The $\theta$ -statistics as defined by Stellingwerf (1978) is first calculated at the frequency $\omega$ , being the frequency of the peak of the power spectrum returned by fasper divided by 2, as the orbital period should normally be twice the period returned from the Fourier analysis because of the primary and secondary eclipse within one orbital period. Then the $\theta$ quantity is calculated for frequencies 0.4 $\omega$ , 0.5 $\omega$ and $\frac{2}{3}\omega$ and the frequency resulting in the lowest $\theta$ - under the condition that it should be smaller than 0.93 $\theta({\omega})$ - is taken as the true orbital frequency. These numbers have been derived from analysis of the objects where the Fourier analysis returned the incorrect orbital period.

The lightcurve is phased with the supposed orbital frequency and from its properties to be described below the candidate EBs are then selected (see Fig. A.1 for an example).

A first limit is set at $m_{\rm lim1} = m_{\rm max} - 0.2\; (m_{\rm max} - m_{\rm min})$ , where $m_{\rm max}$ and $m_{\rm min}$ represent the faintest and brightest magnitude. If there are less than 10 datapoints fainter than $m_{\rm lim1}$ then it is brightened in steps of 0.04 mag until there are.

Then, for a grid of phases $\phi_n$ (N phase steps of 0.01) it is counted how many of the points fainter than $m_{\rm lim1}$ are within 0.05 in phase of $\phi_n$ . The average of this number over the N phase bins defines a "background'' level bkg, and the phase where his number is largest - $n_{\rm max}$ - is recorded, $\phi_{\rm max}$ .

A second limit is set at $m_{\rm lim2} = m_{\rm lim1} - 0.2$ . From all, n₂, phase points fainter than $m_{\rm lim2}$ and within 0.2 in phase of $\phi_{\rm max}$ , a second parameter rms is determined as the rms of the phase difference between $\phi_{\rm max}$ and the phase of the n₂ data points. A third parameter rat is defined as $n_2 / n_{\rm max}$ .

An object is classified as an EB candidate if:

n₂ / bkg > 4.91 and rms <0.0373 and rat <2.01, or
n₂ / bkg > 5.45 and rms <0.036, or
n₂ / bkg > 5.50 and rat <1.64, or
rms <0.022 and rat <2.34.

The principal idea behind this procedure is that compared to the phased light curve of a pulsational variable the one of an EB has at least one minimum, and that the spread in phase at this minimum is limited, and that this also holds at the second magnitude limit (because of the limited width of the eclipse).

The exact value of the parameters have been derived from analysis of a test data set, as described in the main text, and with that setting the program correctly selected 137 out of a 141 objects (=97%).

Appendix B: Newly discovered eclipsing binaries in the Magellanic Clouds

This appendix lists the names and periods (Table B.1) and phased lightcureves (Fig. B.1) of the newly identified EBs in the SMC and LMC.

Table B.1: Properties of newly discovered eclipsing binaries in the Magellanic Clouds.

$\begin{figure} \par\includegraphics[width=17.4cm,clip]{h2886_FigB1_COMPLETE1.eps}\end{figure}$

Figure B.1: Eclipsing binaries. Phased lightcurves of newly discovered eclipsing binaries in the Magellanic Clouds. Phase zero is given by the Julian Date listed in Table B.1.

$\begin{figure} \par\includegraphics[width=17.4cm,clip]{h2886_FigB1_COMPLETE2.eps}\end{figure}$

Figure B.1: continued.

$\begin{figure} \par\includegraphics[width=17.4cm,clip]{h2886_FigB1_COMPLETE3.eps}\end{figure}$

Figure B.1: continued.

$\begin{figure} \par\includegraphics[width=17.4cm,clip]{h2886_FigB1_COMPLETE4.eps}\end{figure}$

Figure B.1: continued.

$\begin{figure} \par\includegraphics[width=17.4cm,clip]{h2886_FigB1_COMPLETE5.eps}\end{figure}$

Figure B.1: continued.

$\begin{figure} \par\includegraphics[width=17.4cm,clip]{h2886_FigB1_COMPLETE6.eps}\end{figure}$

Figure B.1: continued.

$\begin{figure} \par\includegraphics[width=17.4cm,clip]{h2886_FigB1_COMPLETE7.eps}\end{figure}$

Figure B.1: continued.

$\begin{figure} \par\includegraphics[width=17.4cm,clip]{h2886_FigB1_COMPLETE8.eps}\end{figure}$

Figure B.1: continued.

$\begin{figure} \par\includegraphics[width=17.4cm,clip]{h2886_FigB1_COMPLETE9.eps}\end{figure}$

Figure B.1: continued.

$\begin{figure} \par\includegraphics[width=17.4cm,clip]{h2886_FigB1_COMPLETE10.eps}\end{figure}$

Figure B.1: continued.

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE1.eps} \end{figure}$	Figure 1: First 15 entries of the electronically available figure with all phased lightcurves of eclipsing binaries in the direction of the Galactic Bulge. Crosses indicate data points not included in the analysis. Phase zero is given by the Julian Date listed in Table 2.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE2.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE3.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE4.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE5.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE6.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE7.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE8.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE9.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE10.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE11.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE12.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE13.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE14.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE15.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE16.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE17.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE18.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE19.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE20.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE21.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE22.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE23.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE24.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE25.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE26.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE27.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE28.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE29.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE30.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE31.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE32.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE33.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE34.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE35.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE36.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE37.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE38.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE39.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE40.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE41.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE42.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE43.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE44.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE45.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE46.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE47.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE48.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE49.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE50.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE51.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE52.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE53.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE54.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE55.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE56.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE57.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE58.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE59.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE60.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE61.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE62.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE63.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE64.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE65.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE66.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE67.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE68.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE69.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE70.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE71.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE72.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE73.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE74.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE75.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE76.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE77.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE78.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE79.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE80.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE81.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE82.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE83.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE84.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE85.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE86.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE87.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE88.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE89.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE90.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE91.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE92.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE93.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE94.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE95.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE96.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE97.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE98.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE99.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE100.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE101.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE102.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE103.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE104.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE105.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE106.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE107.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE108.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE109.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE110.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE111.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE112.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE113.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE114.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE115.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE116.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE117.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE118.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE119.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE120.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE121.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE122.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE123.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE124.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE125.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE126.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE127.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE128.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE129.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE130.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE131.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE132.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE133.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE134.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE135.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE136.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE137.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE138.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE139.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE140.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE141.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE142.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE143.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE144.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE145.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

$\begin{figure} \par\includegraphics[width=17.3cm,clip]{h2886_Fig1_COMPLETE146.eps}\end{figure}$	Figure 1: continued.
Open with DEXTER

Eclipsing binaries in the Galactic bulge: candidates for distance estimates,

Online Material

Appendix A: The model

Appendix B: Newly discovered eclipsing binaries in the Magellanic Clouds

Eclipsing binaries in the Galactic bulge: candidates for distance estimates^,