Online material

Appendix A: Photometry

Here we provide detailed comments on the photometric observations used in this study and the way we treated them.

• 1.

  Hvar observations were secured in 2007, 2008, and2009 and reduced and transformed into the standardUBV system with the HEC22 release16.1 reduction program via nonlinear transfor-mation formulæ (Harmanecet al. 1994; Harmanec& Horn 1998). This recent version of the programallows modelling of variable extinction during the observingnight, which significantly improves the accuracy of the obser-vations. The typical rms errors of the multinight fit to all standardstars used to define the transformation formulæ ina given observating season are 0ṃ008 for V and B, and 0ṃ010 for U. This is similar for observations from other observing stations reduced with the help of HEC22.

• 2.

  San Pedro Mártir observations were collected during the observational runs in 2001–2003 and 2007. Observations were reduced and transformed into the standard UBV system with the HEC22 release 14.1 reduction program via nonlinear transformation formulæ (Harmanec et al. 1994; Harmanec & Horn 1998). At that station V2368 Oph was used in 2001 as a recommended comparison star for the eclipsing binary U Oph and its magnitude difference relative to several comparison stars (HD 183324, HD 187458, HD 161132, HD 153808, and HD 144206) was derived. For all these stars, save HD 183324, the magnitudes and colours are well established from the calibrated Hvar all-sky photometry. HD 183324 = V1431 Aql was found to be a small-amplitude λ Bootis variable (Kuschnig et al. 1994). It served for a long time as a comparison star for observations of V923 Aql and V1294 Aql in the Photometry of the Bright Northern Be Star Programme (Harmanec et al. 1982, 1994; Pavlovski et al. 1997; Harmanec et al. 1997b) and its variability on longer time scales is safely excluded by numerous Hvar observations. The mean all-sky UBV magnitudes of HD 183324 are accurately derived. It was actually used as a comparison for V2368 Oph only on the night JD 2 452 065, and we feel that its 2 mmag microvariability is not critical for the purpose of this study. For 13 observations secured on JD 2 452 071 (when we recorded the first eclipse of V2368 Oph), it was not possible to derive the differential values for it so we adopted its all-sky values instead, since enough standard stars had been observed during the night, and the nightly transformation coefficients (extinction and its variations and the zero points) could be derived. As soon as we realized that V2368 Oph is a variable, its subsequent observations in 2002 and 2003 were carried out differentially, relative to HD 154660 = HR 6361. This A9V star is a visual binary ADS 10347A with a close companion ADS 10347B at 203, which is some 3ṃ35 fainter than ADS 10347A. The 2007 observations were obtained with a larger diaphragm so that the light of the visual component ADS 10347B was recorded with the brighter component ADS 10347A = HD 6361. We carried out dedicated observations at Hvar to derive the total magnitude of both visual components and added this value to the magnitude differences var – comp. from this season. In all other instances, observations were carried out in such a way as to keep ADS 10347B outside the diaphragm.

• 3.

  Hipparcos all-sky H_p broad-band magnitudes secured between 1989 and 1993 (Perryman & ESA 1997) were transformed to the standard Johnson V magnitude with the nonlinear transformation formula derived by Harmanec (1998). The rms error of the fit per 1 observations is 0ṃ0067. For the solutions, the transmission and the limb darkening coefficients for the H_p passband were considered, however. All data with error flags larger than 1 and one deviating point at HJD 2448661.4682 were omitted.

• 4.

  TNO (Tubitak National Observatory) observations were secured during two consecutive nights in 2003 and were reduced and transformed into the standard UBV system with the HEC22 release 14.1 reduction program via nonlinear transformation formulæ (Harmanec et al. 1994; Harmanec & Horn 1998).

• 5.

  ASAS V magnitude observations were extracted from the public ASAS database (Pojmanski 2002); we used the data from the diaphragm, which gives the smallest rms errors and omitted a few clearly deviating data points.

The journal of all photometric observations is in Table A.1. Homogenized UBV magnitudes of all comparison and check stars used can be found in Table A.2.

Appendix B: Spectroscopy

The journal of all spectroscopic observations can be found in Table B.1. The individual data files are identified there by letters. Below, we provide a few comments on them.

• File A. CCD spectra of V2368 Oph covering the wavelengthregion 6260–6750 Å. They weresecured with a SITe-005 800 × 2000 CCD detector attached to the medium 0.7-m camera of the coudé focus of the Ondřejov 2.0 m telescope (OND). The spectra were obtained between June 2007 and June 2008 and have a linear dispersion of 17.2 Å mm^-1 (red) and a 2-pixel resolving power of about 12600 (11 − 12 km s^-1  per pixel). Their S/N ranges from 50 (1 spectrum) to 370, and the majority have S/N over 200.

• File B. CCD spectra covering the wavelength region 6150–6750 Å with a resolution of 6.6 km s^-1 per pixel. They were obtained at the DAO 1.22-m telescope between August 2007 and September 2009 and have a reciprocal linear dispersion of 10 Å mm^-1. The detector used was a SITe-4 4096 × 2048 CCD, and the 3-pixel resolving power was about 15 000. Their S/N ranges from 100 to 370.

• File C. CCD echelle spectra secured with the 2.14-m reflector of the SPM observatory in April 2003. The CCD detector has 1024 × 1024 pixels, and the setting used covered the wavelength region from 4000 to 6700 Å in grating orders 33 to 60. The nominal resolution of the spectrograph is 18 000 at 5000 Å, which translates into 2-pixel resolution of about 17 km s^-1.

• File D. Another set of echelle CCD spectra from SPM, secured in April 2007.

• File E. The third set of echelle CCD spectra from SPM, secured from May 30 to June 1, 2007. The S/N of the SPM spectra ranges between 120 and 500 for the red, and 85 to 230 for the blue parts of the spectra.

The initial reductions of the DAO spectra (bias subtraction, flatfielding and conversion to 1-D images) were carried out by SY in IRAF. The initial reduction of the SPM and OND spectra was carried out by PE and by Dr. M. Šlechta, respectively, including the wavelength calibration. The remaining reductions of all spectra (including wavelength calibration for the DAO spectra, continuum rectification, and removal of cosmics and flaws) was carried out with the program SPEFO (Horn et al. 1996; Škoda 1996).

© ESO, 2011